File size: 111,266 Bytes
2e1a095 6d5a99d 2e1a095 6d5a99d 2e1a095 6d5a99d 2e1a095 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 | from __future__ import annotations
import argparse
import json
import re
from dataclasses import asdict, dataclass
from pathlib import Path
ROOT_DIR = Path(__file__).resolve().parent.parent
DEFAULT_REPORT = ROOT_DIR / "outputs" / "research-watchlist.md"
DEFAULT_RECOMMENDATION_REPORT = ROOT_DIR / "outputs" / "recommended-free-stack.md"
DEFAULT_DECISION_CARD_JSON = ROOT_DIR / "outputs" / "recommended-decision-card.json"
DEFAULT_DECISION_CARD_REPORT = ROOT_DIR / "outputs" / "recommended-decision-card.md"
@dataclass(frozen=True)
class Candidate:
name: str
kind: str
role: str
status: str
license: str
source: str
hardware: str
why: str
next_step: str
command: str
PERMISSIVE_LICENSE_MARKERS = ("apache-2.0", "mit", "openrail", "open paddleocr ecosystem")
BENCHMARK_ONLY_LICENSE_MARKERS = (
"cc-by-nc",
"non-commercial",
"llama",
"gemma",
"check model card",
"not established",
"fair non-commercial",
)
DEFAULT_SAFE_STATUSES = {"wired default local voice", "wired optional sidecar"}
def is_license_default_safe(candidate: Candidate) -> bool:
license_text = candidate.license.lower()
if "apache-2.0 for specialized msa" in license_text:
return True
if any(marker in license_text for marker in BENCHMARK_ONLY_LICENSE_MARKERS):
return False
return any(marker in license_text for marker in PERMISSIVE_LICENSE_MARKERS)
def is_production_default_safe(candidate: Candidate) -> bool:
status_text = candidate.status.lower()
if "not " in status_text or "caution" in status_text or "correction can alter" in status_text:
return False
return candidate.status in DEFAULT_SAFE_STATUSES and is_license_default_safe(candidate)
def license_policy_violations(candidates: list[Candidate]) -> list[str]:
violations: list[str] = []
for candidate in candidates:
if candidate.status in DEFAULT_SAFE_STATUSES and not is_license_default_safe(candidate):
violations.append(
f"{candidate.name} is marked {candidate.status!r} but has benchmark-only license {candidate.license!r}"
)
if (
not is_production_default_safe(candidate)
and "default" in candidate.next_step.lower()
and not any(marker in candidate.next_step.lower() for marker in ["keep", "do not", "unless"])
):
violations.append(f"{candidate.name} benchmark guidance may imply default promotion too early")
return violations
CANDIDATES = [
Candidate(
name="QARI-OCR 0.4",
kind="ocr",
role="Arabic-book OCR upgrade",
status="wired optional sidecar",
license="Apache-2.0",
source="https://huggingface.co/NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct",
hardware="GPU or strong CPU worker; 4B VLM",
why="Directly trained for Arabic OCR on Islamic books and Arabic manuscripts.",
next_step="Install the sidecar or build the worker with INSTALL_QARI_OCR=1, then benchmark against arabic-max/arabic/arabic-qwen-ocr/katib-ocr/paddleocr/tesseract on the 5-page sample.",
command=(
"python scripts\\benchmark_ocr.py C:\\path\\to\\book-best-5-pages.pdf "
"--page-limit 1 --engines arabic-max arabic arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract"
),
),
Candidate(
name="QARI-OCR 0.4 GGUF",
kind="ocr",
role="portable QARI packaging benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0 via QARI 0.4 model card; confirm GGUF packaging metadata before production",
source="https://huggingface.co/marwan-osama/Qari-OCR-0.4.0-VL-4B-Instruct-GGUF",
hardware="strong worker or local llama.cpp/GGUF runner; QARI 0.4 packaging",
why="A newer GGUF packaging of QARI-OCR 0.4 could make the strongest Arabic-book OCR candidate easier to test in portable runtimes, but it must prove identical or better text quality and acceptable memory/runtime before app wiring.",
next_step="Benchmark the GGUF package externally on the same exported Arabic pages against the wired QARI sidecar, KATIB, Arabic-Qwen, Baseer, PaddleOCR, and Tesseract before considering a llama.cpp-style worker path.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate qari-gguf=outputs\\external-ocr-sample\\qari-gguf.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="PaddleOCR-VL-1.6",
kind="ocr",
role="general document parser upgrade",
status="wired optional sidecar",
license="Apache-2.0",
source="https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.6",
hardware="strong worker; 0.9B document VLM",
why="Fresh Apache-2.0 PaddleOCR document parser release with a June 2026 paper signal; the model card claims SOTA document parsing/text performance and the license file is Apache-2.0, but Arabic-book quality still needs same-page scoring.",
next_step="Build with INSTALL_PADDLEOCR_VL=1 only after the smaller Arabic OCR stack is not clean enough, then benchmark the same 5-page Arabic sample before any full-book run.",
command=(
"python scripts\\benchmark_ocr.py C:\\path\\to\\book-best-5-pages.pdf "
"--page-limit 1 --engines arabic-max arabic arabic-qwen-ocr katib-ocr qari-ocr paddleocr paddleocr-vl tesseract"
),
),
Candidate(
name="oi-OCR",
kind="ocr",
role="Arabic-capable document parser benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/oi-uae/oi-OCR",
hardware="strong worker or external benchmark; PDF/document parser",
why="Apache-2.0 document parser tagged for English/Arabic PDF OCR and layout extraction, with April 2026 ParseBench claims; promising for structured Arabic PDFs but not Arabic-book-specific.",
next_step="Export the same selected Arabic page images and compare its Markdown/text output against QARI/KATIB/Arabic-Qwen/PaddleOCR/Tesseract before considering any wiring.",
command="python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf --out-dir outputs\\external-ocr-sample",
),
Candidate(
name="NuExtract3",
kind="ocr",
role="multilingual document-to-Markdown OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/numind/NuExtract3",
hardware="GPU/large worker; 4B Qwen3.5 VLM, vLLM path recommended",
why="Apache-2.0 4B document understanding model for OCR, document-to-Markdown, tables, forms, invoices, contracts, and multilingual documents. It is not Arabic-book-specific, so compare it externally on the same page images before considering any worker wiring.",
next_step="Use document-to-Markdown/content mode on the exported page images and score the resulting Arabic text against QARI/KATIB/Arabic-Qwen/Baseer/PaddleOCR/Tesseract before promotion.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate nuextract3=outputs\\external-ocr-sample\\nuextract3.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Qianfan-OCR",
kind="ocr",
role="large multilingual document-intelligence OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/baidu/Qianfan-OCR",
hardware="GPU/large worker; 5B BF16 VLM, no hosted inference provider",
why=(
"Apache-2.0 unified end-to-end document OCR/VLM with multilingual/document-intelligence tags "
"and strong general document benchmark claims. It is not Arabic-book-specific and is heavier "
"than the normal free worker path."
),
next_step=(
"Benchmark externally only after QARI/KATIB/Arabic-Qwen/Baseer/PaddleOCR are not clean enough; "
"score it on the same exported Arabic book pages before considering any worker wiring."
),
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"# Run baidu/Qianfan-OCR externally on the same page images, then save merged text as "
"outputs\\external-ocr-sample\\qianfan-ocr.txt.\n"
"python scripts\\score_external_ocr.py --candidate qianfan-ocr=outputs\\external-ocr-sample\\qianfan-ocr.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Chandra OCR 2",
kind="ocr",
role="multilingual structured document OCR benchmark",
status="not default; OpenRAIL weights and GPU-heavy",
license="Apache-2.0 code; modified OpenRAIL-M model weights",
source="https://github.com/datalab-to/chandra",
hardware="GPU/large worker; 4B OCR model, vLLM path recommended",
why="Recent Chandra 2 release supports Arabic among 90+ languages and reports strong structured document OCR/layout performance, but the weights are modified OpenRAIL-M and it is not Arabic-book-specific.",
next_step="Benchmark externally on the same exported page images for hard layouts, tables, forms, or mixed-language pages; keep QARI/KATIB/Arabic-Qwen/Baseer first for Arabic books unless Chandra wins same-page scoring and the license/runtime fit.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"chandra outputs\\external-ocr-sample\\images outputs\\external-ocr-sample\\chandra --method hf\n"
"python scripts\\score_external_ocr.py --candidate chandra=outputs\\external-ocr-sample\\chandra.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="dots.ocr",
kind="ocr",
role="compact multilingual document-layout OCR benchmark",
status="not wired; benchmark externally first",
license="MIT",
source="https://huggingface.co/rednote-hilab/dots.ocr",
hardware="GPU/strong worker; compact 1.7B document parser, vLLM path recommended",
why="MIT multilingual document parser that unifies layout detection and content recognition with reading-order, table, and formula support. It is not Arabic-book-specific, but it is worth scoring on difficult Arabic layouts because public sources mention Arabic/low-resource multilingual OCR.",
next_step="Run externally on the same exported Arabic page images and score the resulting text against QARI/KATIB/Arabic-Qwen/Baseer/PaddleOCR/Tesseract before considering any worker wiring.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate dots-ocr=outputs\\external-ocr-sample\\dots-ocr.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="olmOCR Arabic LoRA v2",
kind="ocr",
role="full-page Arabic manuscript OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0 adapter; confirm base model license/runtime before production",
source="https://huggingface.co/hastyle/olmOCR-arabic-lora-v2",
hardware="GPU/large worker; 7B olmOCR/Qwen2.5-VL base plus LoRA adapter",
why="Apache-2.0 Arabic manuscript OCR LoRA trained on full-page manuscript images, which makes it relevant when line-cropping is impractical; still too heavy for the default free family worker.",
next_step="Run externally on the same exported full-page manuscript images and compare against Ketaba, QARI, HAFITH/Glimpse line workflows, Kraken/eScriptorium, and the wired Arabic OCR baseline before considering any sidecar work.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-manuscript.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"# Run hastyle/olmOCR-arabic-lora-v2 with allenai/olmOCR-2-7B-1025 on the exported full-page images, "
"then save merged text as outputs\\external-ocr-sample\\olmocr-arabic-lora.txt.\n"
"python scripts\\score_external_ocr.py --candidate olmocr-arabic-lora=outputs\\external-ocr-sample\\olmocr-arabic-lora.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Arabic Large Nougat",
kind="ocr",
role="Arabic book page OCR-to-Markdown benchmark",
status="not default; GPL-3.0 and benchmark externally first",
license="GPL-3.0",
source="https://huggingface.co/MohamedRashad/arabic-large-nougat",
hardware="GPU/strong worker; 0.4B VisionEncoderDecoder/Nougat-style model",
why="Arabic-specific OCR model for converting Arabic book page images into structured text/Markdown. It is lighter than many VLM OCR options and directly relevant to book pages, but GPL-3.0 keeps it out of the default public hosted worker.",
next_step="Run externally on the same exported Arabic book page images and compare Markdown/text output against QARI, KATIB, Arabic-Qwen, Baseer, PaddleOCR, Tesseract, and the other external OCR benchmarks before considering any separate license-aware workflow.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"# Run MohamedRashad/arabic-large-nougat externally on the exported page images, "
"then save merged text/Markdown as outputs\\external-ocr-sample\\arabic-large-nougat.txt.\n"
"python scripts\\score_external_ocr.py --candidate arabic-large-nougat=outputs\\external-ocr-sample\\arabic-large-nougat.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="DocTR Arabic FAST/PARSEQ",
kind="ocr",
role="classic Arabic OCR detector/recognizer benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0 detector; recognition card lacks clear metadata, confirm before production",
source="https://huggingface.co/madskills/doctr-fast_base-arabic",
hardware="CPU/GPU worker; DocTR detector plus Arabic PARSEQ recognizer",
why="Apache-2.0 Arabic FAST text detector designed to pair with an Arabic PARSEQ recognizer; useful as a non-VLM Arabic OCR benchmark when large document VLMs are too heavy.",
next_step="Benchmark externally on the same exported Arabic page images and promote only if the recognition model license is confirmed and it beats PaddleOCR/Tesseract/EasyOCR on book text ordering and word preservation.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate doctr-arabic=outputs\\external-ocr-sample\\doctr-arabic.txt "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Kraken/eScriptorium Arabic script",
kind="ocr",
role="open-source historical Arabic-script OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0 engine; model license depends on selected Kraken model",
source="https://kraken.re/main/index.html",
hardware="CPU/GPU external benchmark; best with line segmentation and Arabic-script models",
why="Kraken is an open-source ATR/OCR system optimized for historical and non-Latin scripts. eScriptorium builds on Kraken for manuscript and archival OCR, including Arabic-script workflows, so it is useful when book scans look more like historical print or manuscripts than modern PDFs.",
next_step="Export the same selected page images, run Kraken/eScriptorium with an Arabic-script recognition model or line-cropped workflow, then score the resulting text against the wired Arabic OCR stack before considering any sidecar work.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"kraken -i outputs\\external-ocr-sample\\images\\page-0001.png outputs\\external-ocr-sample\\kraken-page-0001.txt "
"segment ocr -m C:\\path\\to\\arabic-script-kraken-model.mlmodel\n"
"python scripts\\score_external_ocr.py --candidate kraken=outputs\\external-ocr-sample\\kraken.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Kairawan/Qalamus manuscript OCR",
kind="ocr",
role="free Arabic manuscript OCR service signal",
status="not wired; service-only benchmark signal",
license="free web service; engine/package license not established",
source="https://kairawan.org/",
hardware="external web service; not a reusable local worker package",
why=(
"Kairawan is a 2026 Arabic and Islamic manuscript index that offers free automatic "
"Arabic transcription where rights allow, and related public pages mention Qalamus as the "
"manuscript HTR/OCR engine. This is useful evidence that Arabic manuscript OCR is improving, "
"but it is not a drop-in free software dependency for this Vercel plus worker app."
),
next_step=(
"Use only as an external comparison when the source PDF is manuscript-like; do not wire it "
"into the app unless a reusable open engine, API terms, privacy story, and same-page scoring "
"beat QARI/KATIB/Kraken/HAFITH on the selected sample."
),
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-manuscript.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"# If Kairawan/Qalamus output is available for the same pages, save it as "
"outputs\\external-ocr-sample\\kairawan-qalamus.txt.\n"
"python scripts\\score_external_ocr.py --candidate kairawan-qalamus=outputs\\external-ocr-sample\\kairawan-qalamus.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="GLM-OCR Arabic/French documents",
kind="ocr",
role="Arabic administrative/document OCR benchmark",
status="not wired; benchmark externally first",
license="check model card/base license before production use",
source="https://huggingface.co/maloukafer/GLM-OCR-finetuned-documents",
hardware="GPU/strong worker; GLM-OCR LoRA fine-tune",
why="Recent GLM-OCR LoRA fine-tune on thousands of manually annotated Arabic and French scanned documents; useful for forms, receipts, newspapers, and official documents, but not Arabic-book-specific.",
next_step="Benchmark externally for administrative/form-like Arabic PDFs and compare against Arabic-GLM-OCR-v2, QARI, KATIB, Baseer, PaddleOCR, and Tesseract before wiring.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate glm-docs=outputs\\external-ocr-sample\\glm-docs.txt "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="mimoha Arabic OCR",
kind="ocr",
role="minimal Arabic OCR model-card watchlist",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/mimoha/ocr",
hardware="external benchmark; sparse card and no inference provider",
why="Apache-2.0 Arabic OCR model card says it extracts Arabic text from images, but the public card is sparse and references a Mistral OCR-style API path rather than a detailed local runner.",
next_step="Keep as a low-priority external benchmark only; test on exported page images if the stronger Arabic OCR candidates fail or if a clean local runner becomes available.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate mimoha-ocr=outputs\\external-ocr-sample\\mimoha-ocr.txt "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="KATIB 0.8B",
kind="ocr",
role="light Arabic-trained OCR VLM",
status="wired optional sidecar",
license="Apache-2.0",
source="https://huggingface.co/oddadmix/Katib-Qwen3.5-0.8B-0.1",
hardware="smaller worker than QARI; 0.8B VLM",
why="Fine-tuned specifically for Arabic OCR, including printed and handwritten text, while being much smaller than QARI 4B.",
next_step="Install the sidecar or build the worker with INSTALL_KATIB_OCR=1, then benchmark against arabic-max/arabic/arabic-qwen-ocr/qari-ocr/paddleocr/tesseract on the 5-page sample.",
command=(
"python scripts\\benchmark_ocr.py C:\\path\\to\\book-best-5-pages.pdf "
"--page-limit 1 --engines arabic-max arabic arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract"
),
),
Candidate(
name="Arabic-GLM-OCR-v2",
kind="ocr",
role="recent Arabic OCR VLM sidecar",
status="wired optional sidecar",
license="Apache-2.0",
source="https://huggingface.co/sherif1313/Arabic-GLM-OCR-v2",
hardware="strong worker; GLM-OCR fine-tune",
why="Recent Arabic OCR model card claims strong Arabic document extraction and noise reduction; it is wired as an optional sidecar so it can be scored against QARI/KATIB/Arabic-Qwen/Baseer on the target book pages.",
next_step="Install the sidecar or build the worker with INSTALL_ARABIC_GLM_OCR=1, then benchmark it on the same 5-page sample before any full-book run.",
command=(
".\\scripts\\setup_arabic_glm_ocr.ps1\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic arabic-glm-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract"
),
),
Candidate(
name="Arabic-Qwen3.5-OCR-v4",
kind="ocr",
role="small Arabic OCR VLM sidecar",
status="wired optional sidecar",
license="Apache-2.0",
source="https://huggingface.co/sherif1313/Arabic-Qwen3.5-OCR-v4",
hardware="strong worker; 0.9B OCR VLM",
why="Recent Arabic OCR model card claims Arabic printed, handwritten, classical, and diacritic handling in a smaller 0.9B model.",
next_step="Install the sidecar or build the worker with INSTALL_ARABIC_QWEN_OCR=1, then benchmark against arabic-max/arabic/katib-ocr/qari-ocr/paddleocr/tesseract on the 5-page sample.",
command=(
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract"
),
),
Candidate(
name="aNS Qwen3-VL Arabic OCR v3",
kind="ocr",
role="fresh Qwen3-VL Arabic OCR benchmark candidate",
status="not wired; benchmark externally first",
license="check model card/base license before production use",
source="https://huggingface.co/aNS2024/qwen3-vl-arabic-ocr-v3",
hardware="GPU/strong worker; Qwen3-VL-2B fine-tune",
why="Fresh Hugging Face Arabic OCR fine-tune from Qwen3-VL-2B; the public card shows no inference provider and sparse OCR evaluation detail, so it should be scored on the same selected Arabic pages before any app wiring.",
next_step="Export the same 5-page Arabic sample images, run this model externally, and promote only if it beats QARI/KATIB/Arabic-Qwen/Baseer on faithful Arabic text, reading order, runtime, and license fit.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate ans-qwen3-vl-arabic-ocr-v3=outputs\\external-ocr-sample\\ans-qwen3-vl-arabic-ocr-v3.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md "
"--write-json outputs\\external-ocr-sample\\external-ocr-score.json\n"
"python scripts\\model_promotion_gate.py --candidate-name \"aNS Qwen3-VL Arabic OCR v3\" "
"--kind ocr --license \"check model card/base license before production use\" "
"--score-json outputs\\external-ocr-sample\\external-ocr-score.json --same-sample "
"--runtime-ok --privacy-ok --human-reviewed --write-report outputs\\external-ocr-sample\\ocr-promotion-gate.md"
),
),
Candidate(
name="Waraqon v3 Arabic OCR HTML Qari",
kind="ocr",
role="Qari-family structured HTML OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/FatimahEmadEldin/Waraqon-v3-Arabic-OCR-HTML-Qari",
hardware="GPU/strong worker; Qwen2-VL/Qari LoRA path",
why="Apache-2.0 Qari-family fine-tune for Arabic OCR with HTML structure, trained on a Qari 0.3 markdown/HTML-style dataset. It may help complex pages, but audiobook generation needs faithful continuous Arabic text, so same-page text scoring is required before any app wiring.",
next_step="Export the same selected Arabic page images, run Waraqon externally, strip or normalize HTML to readable Arabic text, and promote only if text order, word preservation, runtime, and license fit beat QARI/KATIB/Arabic-Qwen/Baseer.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"# Run FatimahEmadEldin/Waraqon-v3-Arabic-OCR-HTML-Qari externally on the same page images, "
"then save normalized readable text as outputs\\external-ocr-sample\\waraqon-v3.txt.\n"
"python scripts\\score_external_ocr.py --candidate waraqon-v3=outputs\\external-ocr-sample\\waraqon-v3.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md "
"--write-json outputs\\external-ocr-sample\\external-ocr-score.json\n"
"python scripts\\model_promotion_gate.py --candidate-name \"Waraqon v3 Arabic OCR HTML Qari\" "
"--kind ocr --license \"Apache-2.0\" --score-json outputs\\external-ocr-sample\\external-ocr-score.json "
"--same-sample --runtime-ok --privacy-ok --human-reviewed --write-report outputs\\external-ocr-sample\\ocr-promotion-gate.md"
),
),
Candidate(
name="DeepSeek-OCR-2",
kind="ocr",
role="latest general document OCR VLM benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/deepseek-ai/DeepSeek-OCR-2",
hardware="GPU/large worker; 3B DeepSeek OCR VLM",
why="Official 2026 DeepSeek OCR successor with Apache-2.0 licensing, vLLM/SGLang guidance, and public document OCR benchmark results; not Arabic-specific, so score it on the target Arabic book pages before promotion.",
next_step="Benchmark externally on the same exported page images and promote only if it beats QARI/KATIB/Arabic-Qwen/Baseer while the worker can handle 3B VLM inference.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate deepseek-ocr-2=outputs\\external-ocr-sample\\deepseek-ocr-2.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="DeepSeek Arabic OCR v6",
kind="ocr",
role="DeepSeek-OCR Arabic benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/melsiddieg/deepseek_ocr_arabic_v6",
hardware="GPU/large worker; 3B DeepSeek-OCR fine-tune",
why="Newer Apache-2.0 Arabic-labeled DeepSeek-OCR fine-tune than v4/v5, but the model card has sparse evaluation detail and no hosted inference provider, so keep it as an external same-page benchmark.",
next_step="Benchmark externally on the same exported page images and promote only if it beats the wired Arabic OCR stack on the target pages.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate deepseek-ar-v6=outputs\\external-ocr-sample\\deepseek-ar-v6.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Loay Arabic-OCR-DeepSeek-OCR-2",
kind="ocr",
role="Arabic DeepSeek-OCR-2 layout benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/loay/Arabic-OCR-DeepSeek-OCR-2",
hardware="GPU/large worker; 3B DeepSeek-OCR-2 fine-tune",
why="Recent Apache-2.0 merged DeepSeek-OCR-2 Arabic fine-tune aimed at high-precision OCR and structural layout analysis from Arabic images; promising for pages where layout preservation matters, but it is still 3B-class and must beat the Arabic-specific QARI/KATIB/Arabic-Qwen/Baseer stack on the same book pages.",
next_step="Benchmark externally on the same exported Arabic page images and promote only if it beats QARI/KATIB/Arabic-Qwen/Baseer without hallucinating or reordering text and the worker can handle 3B VLM inference.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate loay-deepseek-ocr-2=outputs\\external-ocr-sample\\loay-deepseek-ocr-2.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Arabic-English handwritten OCR v3",
kind="ocr",
role="handwriting/manuscript OCR benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/sherif1313/Arabic-English-handwritten-OCR-v3",
hardware="strong worker; Qwen2.5-VL 3B-class model, 7.5GB model assets",
why="Apache-2.0 Arabic/English handwritten OCR model with model-card claims for handwritten and manuscript pages; useful when the PDF has marginal notes, handwriting, or older script that the book OCR stack misses.",
next_step="Benchmark externally on the same exported page images and promote only if it beats QARI/KATIB/Arabic-Qwen/Baseer on the target handwriting or manuscript pages.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate handwritten-v3=outputs\\external-ocr-sample\\handwritten-v3.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Arabic handwritten OCR 4-bit Qwen2.5-VL",
kind="ocr",
role="lighter quantized handwriting/manuscript OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/sherif1313/Arabic-handwritten-OCR-4bit-Qwen2.5-VL-3B-v3",
hardware="strong worker; 4-bit Qwen2.5-VL 3B-class model, about 2.44GB model assets",
why="Apache-2.0 4-bit Arabic handwritten OCR checkpoint that is much smaller than the full Arabic-English handwritten OCR model; useful when handwriting or manuscript pages matter and a free worker cannot handle the larger checkpoint.",
next_step="Benchmark externally on the same exported handwriting/manuscript page images and promote only if it beats the wired Arabic OCR stack without introducing hallucinated or reordered text.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate handwritten-4bit=outputs\\external-ocr-sample\\handwritten-4bit.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="NAKBA Arabic manuscript line OCR baseline",
kind="ocr",
role="Arabic manuscript line-transcription benchmark",
status="not wired; benchmark externally first",
license="check model card/base license before production use",
source="https://huggingface.co/U4RASD/ar-ms-baseline",
hardware="large GPU/strong worker; Qwen3-VL-8B LoRA baseline, H100 used in reported evaluation",
why="NAKBA NLP 2026 shared-task baseline for Arabic manuscript line transcription, with reported CER/WER on released line-image tests; useful only when the source PDF has manuscript or memoir-style handwritten line images.",
next_step="Keep external and line-level only unless a preprocessing step crops pages into lines; compare with handwritten-v3 and handwritten-4bit on the same manuscript images before any wiring.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-manuscript.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate nakba-ms-line=outputs\\external-ocr-sample\\nakba-ms-line.txt "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="HAFITH",
kind="ocr",
role="historical Arabic manuscript line OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/mdnaseif/hafith",
hardware="GPU/strong worker likely; 642M line OCR model with line segmentation required",
why="Apache-2.0 Arabic-native OCR model for historical manuscript line recognition with 5.10% CER claims, but it operates on pre-segmented text lines rather than full PDF pages.",
next_step="Use only when a scanned book looks like historical print or manuscript pages; crop pages into text lines first, then compare HAFITH line output against Kraken/eScriptorium, NAKBA line OCR, and the wired OCR baseline.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"# Crop selected pages into text-line images before running mdnaseif/hafith, then save the merged line text as "
"outputs\\external-ocr-sample\\hafith.txt.\n"
"python scripts\\score_external_ocr.py --candidate hafith=outputs\\external-ocr-sample\\hafith.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Glimpse RTL OCR",
kind="ocr",
role="Arabic/Persian RTL text-line OCR benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/surfiniaburger/unsloth_finetune_ocr_arabic",
hardware="GPU/strong worker likely; 1B text-line VLM fine-tune",
why="Apache-2.0 Arabic/Persian RTL text-line OCR model fine-tuned from an ERNIE/PaddleOCR-VL path with 6.97% CER claims on unseen RTL text lines, but it is line-level rather than a full-page PDF OCR default.",
next_step="Use only after selected PDF pages are cropped into Arabic/Persian text-line images; compare against HAFITH, NAKBA line OCR, Qwen3-VL Persian/Arabic line OCR, Kraken/eScriptorium, and the wired OCR baseline.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"# Crop selected pages into text-line images before running surfiniaburger/unsloth_finetune_ocr_arabic, then save the merged line text as "
"outputs\\external-ocr-sample\\glimpse-rtl.txt.\n"
"python scripts\\score_external_ocr.py --candidate glimpse-rtl=outputs\\external-ocr-sample\\glimpse-rtl.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Arabic OCR Qwen2.5-VL GGUF",
kind="ocr",
role="QariOCR-trained Arabic/English GGUF OCR benchmark",
status="not wired; benchmark externally first",
license="check model card/base license before production use",
source="https://huggingface.co/mo1998/arabic-ocr-qwen2.5-vl",
hardware="large GPU/strong worker; Qwen2.5-VL 7B-class GGUF/Unsloth path",
why="Arabic/English OCR fine-tune trained on QariOCR v0.3 mixed data and aimed at scanned books, religious texts, handwritten forms, and mixed-language documents; useful as an external GGUF-style comparison against QARI 0.4 and the wired Arabic OCR stack.",
next_step="Benchmark externally on the same exported page images, confirm license fit, and promote only if it beats QARI/KATIB/Arabic-Qwen/Baseer on the target pages while the worker can handle 7B-class inference.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate qwen25-gguf=outputs\\external-ocr-sample\\qwen25-gguf.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Qwen3-VL Persian/Arabic line OCR",
kind="ocr",
role="line-level Persian/Arabic OCR benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/mohajesmaeili/Qwen3-VL-2B-Persian-Arabic-Ocr-v1.0",
hardware="strong worker; Qwen3-VL 2B line-level OCR model",
why="Apache-2.0 Qwen3-VL 2B fine-tune for Persian/Arabic text-line OCR; interesting for cropped line images, but its model card says it was trained on individual text lines and is not designed for full-page OCR.",
next_step="Keep external unless a preprocessing/layout step crops Arabic book pages into lines; benchmark only on exported/cropped line images and compare against the full-page Arabic OCR stack before any wiring.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate qwen3-line-ocr=outputs\\external-ocr-sample\\qwen3-line-ocr.txt "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Loay Arabic-OCR-Qwen2.5-VL-7B",
kind="ocr",
role="large Arabic OCR VLM benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0 via Qwen2.5-VL base, confirm model-card license before production",
source="https://huggingface.co/loay/Arabic-OCR-Qwen2.5-VL-7B-Vision",
hardware="large GPU/strong worker; Qwen2.5-VL 7B fine-tune",
why="Arabic OCR VLM fine-tuned from Qwen2.5-VL-7B for OCR on Arabic text from images; useful as a high-capacity comparison when smaller Arabic OCR sidecars fail.",
next_step="Benchmark externally on the same exported page images and promote only if it beats QARI/KATIB/Arabic-Qwen/Baseer while the worker can handle 7B-class runtime.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate loay-qwen25=outputs\\external-ocr-sample\\loay-qwen25.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="DIMI Arabic OCR v2",
kind="ocr",
role="large Arabic OCR LoRA benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/AhmedZaky1/DIMI-Arabic-OCR-V2",
hardware="large GPU/strong worker; Qwen2.5-VL 7B LoRA with 4-bit option",
why="Recent Arabic OCR model card reports improved diacritics handling, printed-document extraction, and lower WER/CER than its v1 model, but it is 7B-class and needs external scoring on the target book pages.",
next_step="Benchmark externally on the same exported page images and promote only if it beats the wired Arabic OCR sidecars while the worker can handle the 7B LoRA runtime.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic arabic-glm-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate dimi-v2=outputs\\external-ocr-sample\\dimi-v2.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="AtlasOCR",
kind="ocr",
role="Darija/Moroccan Arabic OCR watchlist",
status="not wired; benchmark externally first",
license="check model card before production use",
source="https://huggingface.co/atlasia/AtlasOCR",
hardware="strong worker; Qwen2.5-VL 3B fine-tune",
why="First open-source Darija OCR model; useful if the PDFs contain Moroccan Arabic or Darija rather than standard printed MSA.",
next_step="Benchmark only for Darija/Moroccan content, and confirm licensing before any production wiring.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate atlasocr=outputs\\external-ocr-sample\\atlasocr.txt "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Ketaba-OCR LoRA",
kind="ocr",
role="Arabic manuscript OCR benchmark candidate",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/HassanB4/Ketaba-OCR-LoRA",
hardware="strong worker; Qwen2.5-VL-3B base plus LoRA adapter",
why="Arabic manuscript benchmark winner candidate with reported low character error rate, useful when book/manuscript scans defeat the wired OCR stack.",
next_step="Benchmark externally on one or two page images before considering a sidecar because it needs a separate base VLM plus adapter setup.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate ketaba=outputs\\external-ocr-sample\\ketaba.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Qari-OCR-LoRA",
kind="ocr",
role="Arabic manuscript OCR secondary QARI-family benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/HassanB4/Qari-OCR-LoRA",
hardware="strong worker; QARI/Qwen2-VL-family base plus LoRA adapter",
why="Apache-2.0 experimental QARI-family LoRA from the NakbaNLP 2026 Arabic manuscript OCR task; useful as a secondary manuscript benchmark after Ketaba because its own model card says Ketaba was the primary winning submission.",
next_step="Benchmark externally on the same manuscript-like page images only after QARI 0.4, Ketaba, KATIB, and the wired OCR stack have been scored.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-manuscript.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate qari-ocr-lora=outputs\\external-ocr-sample\\qari-ocr-lora.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Tawkeed OCR",
kind="ocr",
role="Arabic-first OCR sidecar",
status="wired optional sidecar",
license="Apache-2.0",
source="https://huggingface.co/tawkeed-sa/tawkeed-ocr",
hardware="strong worker or edge/GPU test; QARI v0.3 2B fork",
why="Arabic-first OCR model forked from QARI-OCR v0.3 and fine-tuned for Arabic documents, handwriting, and scene text; useful to test when QARI 0.4 is too heavy or when edge-style deployment matters.",
next_step="Install the sidecar or build the worker with INSTALL_TAWKEED_OCR=1, then benchmark against QARI 0.4, KATIB, Arabic-Qwen, Baseer, and Tesseract on the same 5-page sample.",
command=(
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic tawkeed-ocr arabic-qwen-ocr katib-ocr qari-ocr baseer-ocr paddleocr tesseract"
),
),
Candidate(
name="Falcon-OCR",
kind="ocr",
role="compact document OCR VLM watchlist",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/tiiuae/Falcon-OCR",
hardware="strong worker; 300M document VLM",
why="Smaller Apache-2.0 document OCR VLM that may be useful when VLM OCR is needed but larger parsers are too heavy.",
next_step="Run an external one-page image benchmark and compare cleaned Arabic word count, quality score, and reading order against QARI, Arabic-Qwen, and Tesseract.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract --json > "
"outputs\\external-ocr-sample\\wired-ocr-baseline.json\n"
"python scripts\\score_external_ocr.py --candidate falcon=outputs\\external-ocr-sample\\falcon.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="Baseer OCR V1.0",
kind="ocr",
role="Arabic legal/complex document OCR sidecar",
status="wired optional sidecar",
license="Apache-2.0",
source="https://huggingface.co/AbdoTarek/Baseer-OCR-V1.0",
hardware="strong worker; Qwen2-VL 2B fine-tune",
why="Arabic-specific VLM OCR for complex legal documents, multi-column layouts, stamps, tables, and handwritten/printed Arabic.",
next_step="Install the sidecar or build the worker with INSTALL_BASEER_OCR=1, then benchmark against arabic-max/arabic/arabic-qwen-ocr/katib-ocr/qari-ocr/paddleocr/tesseract on the 5-page sample.",
command=(
".\\scripts\\setup_baseer_ocr.ps1\n"
"python scripts\\benchmark_ocr.py test_pdfs\\book-best-5-pages.pdf --page-limit 5 "
"--engines arabic-max arabic baseer-ocr arabic-qwen-ocr katib-ocr qari-ocr paddleocr tesseract"
),
),
Candidate(
name="Raqim post-OCR correction",
kind="ocr",
role="Arabic OCR correction research caution",
status="not wired; correction can alter source text",
license="open-access paper; implementation/license not established for this app",
source="https://www.sciencedirect.com/science/article/pii/S187705092600058X",
hardware="depends on correction model; dictionary plus LLM workflow",
why="2026 research reports Arabic OCR correction gains by combining dictionary-based correction and LLMs, but correction may rewrite exact book/religious wording, so it should not run automatically before TTS.",
next_step="Keep as a manual review/research idea only. If used later, expose it as an explicit reviewed correction mode, never as the default reading path.",
command="python scripts\\dry_run_pdf.py C:\\path\\to\\arabic-book.pdf --include-speech-text --speech-sample-chars 1200",
),
Candidate(
name="Arabic Legal Documents OCR 1.0",
kind="ocr",
role="legal/structured Arabic OCR benchmark candidate",
status="not default; Gemma license and domain-specific",
license="Gemma license",
source="https://huggingface.co/bakrianoo/arabic-legal-documents-ocr-1.0",
hardware="GPU/large worker; Gemma-3-4B-IT VLM fine-tune",
why="Recent Arabic legal-document OCR VLM optimized for low-quality scanned Arabic legal documents and structured extraction; useful only when the target PDF is legal/structured rather than a normal book.",
next_step="Benchmark externally on exported page images only for legal or form-like Arabic PDFs; keep it out of the default family audiobook stack because it is Gemma-licensed and domain-specific.",
command=(
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample\n"
"python scripts\\score_external_ocr.py --candidate legal-docs-ocr=outputs\\external-ocr-sample\\legal-docs-ocr.txt "
"--write-report outputs\\external-ocr-sample\\external-ocr-score.md"
),
),
Candidate(
name="SILMA TTS",
kind="tts",
role="default Arabic neural voice",
status="wired default local voice",
license="MIT code, Apache-2.0 model weights",
source="https://huggingface.co/silma-ai/silma-tts",
hardware="CPU worker practical baseline",
why="Arabic-focused Fusha/MSA voice with normalization and tashkeel options.",
next_step="Keep as the first listen-test voice for every book sample.",
command="python scripts\\benchmark_voices.py --voices silma-local espeak-ar-clear --write-report outputs\\voice-benchmark-report.md",
),
Candidate(
name="Habibi-TTS MSA",
kind="tts",
role="optional Arabic MSA voice comparison",
status="wired optional sidecar",
license="Apache-2.0 for specialized MSA; other variants may be non-commercial",
source="https://github.com/SWivid/Habibi-TTS",
hardware="CPU/GPU worker depending on model/runtime",
why="Arabic-specific 2026 TTS family worth comparing against SILMA on MSA passages.",
next_step="Install the optional sidecar and listen against the same cleaned OCR sample.",
command="python scripts\\benchmark_voices.py --voices silma-local habibi-msa espeak-ar-clear --write-report outputs\\voice-benchmark-report.md",
),
Candidate(
name="Mishkala Tashkeel",
kind="tts",
role="Arabic pronunciation preprocessor benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/flokymind/mishkala",
hardware="CPU/worker practical; 12.5M diacritization model",
why="Lightweight Apache-2.0 Arabic diacritization model with model-card DER claims; useful to test whether adding tashkeel improves SILMA, Habibi, Supertonic, or eSpeak pronunciation on the exact book text.",
next_step="Benchmark on the same cleaned speech sample before wiring. Promote only if listening tests improve pronunciation without changing meaning or adding distracting/incorrect harakat.",
command=(
"python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf "
"--env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample\n"
"# Then diacritize outputs\\external-tts-sample\\arabic-tts-sample.txt with flokymind/mishkala "
"and listen-test the same text with SILMA/Habibi/Supertonic before enabling any automatic path."
),
),
Candidate(
name="Tashkeel-350M",
kind="tts",
role="larger Arabic pronunciation preprocessor benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/Etherll/Tashkeel-350M",
hardware="GPU/strong worker likely; 350M Arabic diacritization model",
why="Apache-2.0 Arabic diacritization model that is much larger than Mishkala. It may improve pronunciation for some voices, but larger tashkeel models can still add wrong harakat, so compare by listening before production use.",
next_step="Export the same cleaned Arabic TTS sample, create a Tashkeel-350M diacritized copy, synthesize plain/Mishkala/Tashkeel-350M with the same voice, and score meaning preservation plus long-listen comfort.",
command=(
"python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env "
"--out-dir outputs\\external-tts-sample\n"
"# Then diacritize outputs\\external-tts-sample\\arabic-tts-sample.txt with Etherll/Tashkeel-350M and "
"score it with scripts\\score_tts_preprocessor.py before enabling any automatic path."
),
),
Candidate(
name="Mushkil",
kind="tts",
role="AraT5V2 Arabic pronunciation preprocessor benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/riotu-lab/mushkil",
hardware="GPU/strong worker likely; AraT5V2 Arabic diacritization model",
why="Apache-2.0 AraT5V2 Arabic diacritization model. It belongs in the same pronunciation-preprocessor benchmark lane as Mishkala and Tashkeel-350M, but automatic harakat can still change perceived meaning or comfort.",
next_step="Export the same cleaned Arabic TTS sample, create a Mushkil-diacritized copy, synthesize plain/Mishkala/Tashkeel-350M/Mushkil with the same voice, and score meaning preservation plus long-listen comfort.",
command=(
"python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env "
"--out-dir outputs\\external-tts-sample\n"
"# Then diacritize outputs\\external-tts-sample\\arabic-tts-sample.txt with riotu-lab/mushkil and "
"score it with scripts\\score_tts_preprocessor.py before enabling any automatic path."
),
),
Candidate(
name="Thaka KSAA-2026 speech diacritization",
kind="tts",
role="Arabic diacritization research signal",
status="research signal only; no deployable model yet",
license="CC BY 4.0 paper; implementation/model license not established",
source="https://arxiv.org/abs/2605.25928",
hardware="not applicable until code/weights are released; paper describes CATT-Whisper ensemble inference",
why="Late-May 2026 winning KSAA shared-task system for Arabic speech dictation diacritization. It is useful evidence that acoustic-plus-text diacritization can improve Arabic pronunciation work, but it is not a plug-in free model for this PDF-to-audio app.",
next_step="Track for released code/weights or a permissive checkpoint. Until then, keep website preprocessing limited to same-sample Mishkala/Tashkeel-350M/Mushkil listening tests and meaning-preservation scoring.",
command=(
"python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env "
"--out-dir outputs\\external-tts-sample\n"
"# Do not wire Thaka automatically; use the paper as a benchmark signal until a permissive implementation exists."
),
),
Candidate(
name="3arab-TTS 500M",
kind="tts",
role="new Arabic-only voice benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/sherif1313/3arab-TTS-500M-v1",
hardware="GPU/strong worker likely; 500M Arabic RF-DiT model",
why="Very recent Arabic-only Apache-2.0 TTS model trained on Arabic speech datasets; promising enough to benchmark against SILMA/Habibi before any app wiring.",
next_step="Export the same cleaned Arabic text used for SILMA/Habibi, then compare base and VoiceDesign variants for audiobook comfort, stability, and long-form pacing.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="KaniTTS Arabic",
kind="tts",
role="high-speed Arabic voice benchmark",
status="not wired; benchmark externally first",
license="model card says Apache-2.0, but Hugging Face metadata reports lfm1.0; confirm before production",
source="https://huggingface.co/nineninesix/kani-tts-400m-ar",
hardware="GPU/strong worker likely; 400M model, reported low latency with vLLM",
why="Recent Arabic-only TTS model with high-speed claims; promising for faster audiobook generation, but current Hugging Face metadata reports lfm1.0 while the card text says Apache-2.0.",
next_step="Export the same cleaned Arabic sample used for SILMA/Habibi, then benchmark naturalness, skipped words, pacing, runtime, and license fit before considering app wiring.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Emirati VITS Male",
kind="tts",
role="Gulf/Emirati Arabic voice benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/vadimbelsky/emirati-vits-male-1.0",
hardware="GPU/NeMo runtime; 22.05 kHz VITS voice",
why="Apache-2.0 bilingual Emirati Arabic/English VITS voice that may be useful for Gulf dialect material, but dialectal tuning makes it a comparison candidate rather than the default MSA book voice.",
next_step="Benchmark only when the target PDF benefits from Emirati/Gulf pronunciation; keep SILMA/Habibi ahead for MSA books unless listening tests say otherwise.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="VoxCPM2",
kind="tts",
role="strong-worker multilingual voice watchlist",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/openbmb/VoxCPM2",
hardware="GPU/large worker; 2B model, about 8GB VRAM guidance",
why="Apache-2.0 multilingual TTS with Arabic support, 48 kHz output, and voice design/cloning.",
next_step="Benchmark externally with the same cleaned Arabic sample before deciding whether it is worth integrating.",
command="python scripts\\dry_run_pdf.py C:\\path\\to\\arabic-book.pdf --include-speech-text --speech-sample-chars 1200",
),
Candidate(
name="Voxtral TTS",
kind="tts",
role="non-commercial Arabic-capable strong-worker voice watchlist",
status="not default; non-commercial and GPU-heavy",
license="CC-BY-NC-4.0",
source="https://huggingface.co/mistralai/Voxtral-4B-TTS-2603",
hardware="GPU/large worker; 4B BF16 model, 16GB+ GPU guidance",
why="Mistral's open-weight TTS model lists Arabic among 9 supported languages and has strong latency/quality claims, but the model card says it inherits CC-BY-NC-4.0 and recommends a large GPU setup.",
next_step="Benchmark only as a personal/non-commercial strong-worker comparison using the same cleaned Arabic sample; do not wire it as the default public/free website voice.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="OmniVoice",
kind="tts",
role="priority permissive strong-worker voice benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/k2-fsa/OmniVoice",
hardware="GPU/large worker; 0.6B model",
why="Apache-2.0 zero-shot TTS with 646-language coverage, Arabic included, high current usage, and published 2026 OmniVoice evidence; benchmark it before heavier or license-unclear voices when a stronger worker is available.",
next_step="Export the same cleaned Arabic text used for SILMA/Habibi and compare Arabic naturalness, speed, and setup complexity before wiring it into the app.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="OmniVoice Arabic LoRA",
kind="tts",
role="Arabic adapter for OmniVoice benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/vivooglobal/omnivoice-lora-ar",
hardware="GPU/large worker; OmniVoice plus LoRA adapter",
why="Arabic LoRA adapter for OmniVoice intended to improve zero-shot Arabic voice cloning quality.",
next_step="Benchmark only after the base OmniVoice command is working, using the exact same cleaned Arabic sample and reference audio.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Arabic-text-to-speech OmniVoice",
kind="tts",
role="Arabic-focused OmniVoice packaging benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/bilalRHCH/Arabic-text-to-speech",
hardware="GPU/large worker; OmniVoice 0.6B packaging with Arabic-focused model card",
why="Apache-2.0 Arabic-labeled OmniVoice package with 646-language OmniVoice support and a demo Space signal; useful as a same-sample Arabic voice benchmark, but still not proven for long MSA audiobook passages.",
next_step="Export the same cleaned Arabic sample used for SILMA/Habibi and compare naturalness, skipped words, repetition, runtime, and setup complexity before any app wiring.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Lahgtna OmniVoice v2",
kind="tts",
role="Arabic dialect OmniVoice benchmark",
status="not wired; benchmark externally first",
license="license not declared on model card",
source="https://huggingface.co/oddadmix/lahgtna-omnivoice-v2",
hardware="GPU/large worker; OmniVoice 0.6B fine-tune",
why="New Arabic dialect TTS model based on OmniVoice, with broad Arabic dialect tags and diacritics support; promising for conversational dialect audio but not the MSA book default.",
next_step="Benchmark externally only when dialect pronunciation matters, confirm licensing before production, and keep SILMA/Habibi ahead for MSA books until listening tests prove otherwise.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="TADA multilingual TTS",
kind="tts",
role="strong-worker off-script-resistant voice benchmark",
status="not default; Llama 3.2 license and GPU-heavy",
license="Llama 3.2 license",
source="https://huggingface.co/HumeAI/tada-3b-ml",
hardware="GPU/large worker; 3B-class multilingual model plus language aligners",
why="Free/open-weight multilingual TTS with Arabic support and a text-acoustic alignment design meant to reduce skipped or invented words, which matters for long book narration, but its Hugging Face metadata reports the Llama 3.2 license rather than Apache-2.0.",
next_step="Export the same cleaned Arabic sample and benchmark with language='ar' only if the Llama 3.2 license is acceptable; keep SILMA/Habibi ahead for the permissive default.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Lahgtna Chatterbox",
kind="tts",
role="Arabic dialect TTS benchmark",
status="not wired; benchmark externally first",
license="MIT",
source="https://huggingface.co/oddadmix/lahgtna-chatterbox-v1",
hardware="GPU recommended; Chatterbox multilingual base",
why="MIT Arabic dialect TTS model covering multiple dialect tags; promising when conversational dialect pronunciation matters more than MSA audiobook narration.",
next_step="Export the same cleaned Arabic text and listen for repetition/stability before considering app wiring.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="NAMAA-Saudi-TTS",
kind="tts",
role="Saudi Arabic Chatterbox dialect benchmark",
status="not wired; benchmark externally first",
license="MIT",
source="https://huggingface.co/NAMAA-Space/NAMAA-Saudi-TTS",
hardware="GPU recommended; 0.5B Chatterbox Multilingual fine-tune",
why="MIT-licensed Saudi Arabic TTS model built on Chatterbox Multilingual; useful for Saudi/Gulf dialect material, but its model card says it targets everyday Saudi speech rather than MSA books.",
next_step="Benchmark only when Saudi dialect pronunciation fits the target PDF; keep SILMA/Habibi first for MSA books and compare against Saudi Arabic Qwen3-TTS and Emirati voices before wiring.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="NAMAA-Egyptian-TTS",
kind="tts",
role="Egyptian Arabic Chatterbox dialect benchmark",
status="not wired; benchmark externally first",
license="MIT",
source="https://huggingface.co/NAMAA-Space/NAMAA-Egyptian-TTS",
hardware="GPU recommended; 0.5B Chatterbox Multilingual fine-tune",
why="MIT-licensed NAMAA Egyptian Arabic TTS model built on Chatterbox Multilingual with local/hosted inference examples and a live demo, but its card says it targets everyday Egyptian speech rather than MSA books and notes number/pronunciation limitations.",
next_step="Benchmark only for Egyptian/dialectal PDFs using the same cleaned Arabic sample; keep SILMA/Habibi first for MSA books and compare against Egyptian Arabic Chatterbox, NileTTS-XTTS, and Egyptian Arabic Qwen3-TTS before any wiring.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Saudi Chatterbox fine-tune",
kind="tts",
role="Saudi Arabic Chatterbox dialect benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/FatimahEmadEldin/saudi-tts-chatterbox-finetuned",
hardware="GPU recommended; Chatterbox Multilingual T3 fine-tune",
why="Apache-2.0 Saudi Arabic Chatterbox Multilingual fine-tune trained on Saudi podcast and number-normalized speech data; useful as a same-family comparison against NAMAA-Saudi-TTS.",
next_step="Benchmark only for Saudi/Gulf dialect fit using the same cleaned Arabic sample; keep SILMA/Habibi first for MSA books and compare against NAMAA-Saudi-TTS, Saudi Qwen3-TTS, and Emirati voices before wiring.",
command=(
"python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf "
"--env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample\n"
"python -c \"# Load ChatterboxMultilingualTTS, apply FatimahEmadEldin/saudi-tts-chatterbox-finetuned "
"T3 weights, synthesize outputs\\external-tts-sample\\arabic-tts-sample.txt, language_id='ar'\""
),
),
Candidate(
name="Saudi TTS",
kind="tts",
role="Saudi Arabic dialect TTS benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/AhmedEladl/saudi-tts",
hardware="GPU/worker benchmark; model card has demo Spaces but no inference-provider deployment",
why=(
"Apache-2.0 high-quality Saudi Arabic dialect voice candidate. It may be useful for Gulf/Saudi "
"material, but it is dialect-specific and not a proven MSA audiobook voice."
),
next_step=(
"Benchmark beside NAMAA-Saudi-TTS, Saudi Chatterbox fine-tune, Saudi Arabic Qwen3-TTS, "
"and SILMA/Habibi on the same cleaned Arabic sample before any app wiring."
),
command=(
"python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf "
"--env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample"
),
),
Candidate(
name="Egyptian Arabic Chatterbox",
kind="tts",
role="Egyptian Arabic dialect Chatterbox benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/AliAbdallah/egyptian-arabic-tts-chatterbox",
hardware="GPU recommended; Chatterbox Multilingual fine-tune",
why="Apache-2.0 Egyptian Arabic Chatterbox fine-tune with 120 hours of clean Egyptian Arabic training data; useful for Egyptian dialect content, but not a general MSA book voice and the model card says GPU is needed for real-time inference.",
next_step="Benchmark only for Egyptian/dialectal PDFs using the same cleaned Arabic sample; keep SILMA/Habibi first for MSA books unless listening tests clearly favor the dialect voice for that text.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="NileTTS-XTTS",
kind="tts",
role="Egyptian Arabic XTTS benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/KickItLikeShika/NileTTS-XTTS",
hardware="GPU recommended; XTTS v2 fine-tune",
why="Apache-2.0 Egyptian Arabic XTTS fine-tune from the 2026 NileTTS work; useful for Egyptian dialect content, but not a general MSA book voice.",
next_step="Benchmark only for Egyptian/dialectal PDFs using the same cleaned Arabic sample; keep SILMA/Habibi first for MSA books unless listening tests clearly favor this dialect voice.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Arabic XTTS-v2 Egyptian fine-tune",
kind="tts",
role="Egyptian Arabic XTTS voice-quality benchmark",
status="not wired; benchmark externally first",
license="MIT project; XTTS-v2 base uses Coqui Public Model License",
source="https://huggingface.co/Moeeldouma/arabic-tts-xtts-v2",
hardware="GPU recommended; XTTS-v2 fine-tune and voice-cloning runtime",
why="Recent Arabic XTTS-v2 improvement project with Egyptian speaker fine-tuning and documented same-text comparisons. It is worth a listening benchmark for dialectal material, but the XTTS-v2 base license and voice-cloning setup keep it out of the permissive default path.",
next_step="Benchmark externally only for Egyptian/dialectal PDFs, confirm CPML fit for the intended personal use, and compare long-form stability against SILMA/Habibi/NileTTS-XTTS on the exact same cleaned text.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Chatterbox-Multilingual",
kind="tts",
role="MIT Arabic-capable multilingual TTS benchmark",
status="not wired; benchmark externally first",
license="MIT",
source="https://github.com/resemble-ai/chatterbox",
hardware="GPU recommended; multilingual TTS/voice cloning",
why="MIT-licensed Chatterbox-Multilingual lists Arabic among 23 supported languages and has local generation examples, making it a strong free Arabic voice benchmark candidate.",
next_step="Export the same cleaned Arabic text used for SILMA/Habibi and compare naturalness, pacing, watermarking, and setup complexity before wiring it into the app.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Chatterbox Arabic fine-tune",
kind="tts",
role="MSA-focused Chatterbox Arabic adapter benchmark",
status="not wired; benchmark externally first",
license="MIT",
source="https://huggingface.co/juliardi/chatterbox-multilingual-finetuned-arabic",
hardware="GPU recommended; Chatterbox Multilingual LoRA/adapter",
why="MIT Chatterbox Arabic fine-tune claims improved Arabic pronunciation, diacritics, intonation, MSA support, and common dialect support, making it worth a same-text listening test against SILMA/Habibi before any app wiring.",
next_step="Export the same cleaned Arabic sample used for SILMA/Habibi and compare MSA pronunciation, skipped words, repetition, long-sentence stability, and runtime; keep it benchmark-only until it wins human listening tests.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Chatterbox-Multilingual ONNX",
kind="tts",
role="CPU/ONNX Arabic-capable voice benchmark",
status="not wired; benchmark externally first",
license="MIT",
source="https://huggingface.co/onnx-community/chatterbox-multilingual-ONNX",
hardware="CPU/ONNX-capable worker; multilingual Chatterbox runtime",
why="MIT ONNX packaging for Chatterbox-Multilingual with Arabic included in the supported language list. It may be easier to benchmark on CPU-style free workers than heavier GPU voice models.",
next_step="Export the same cleaned Arabic text used for SILMA/Habibi and compare ONNX Chatterbox naturalness, repetition, pacing, pronunciation, and runtime before any app wiring.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Spark-TTS Arabic",
kind="tts",
role="classical Arabic voice-cloning benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/azeddinShr/Spark-TTS-Arabic-Complete",
hardware="GPU/large worker; Spark-TTS repo required",
why="Apache-2.0 Arabic fine-tune of Spark-TTS on ClArTTS, aimed at Classical/MSA narration, but the model card says non-diacritized input is out of scope and the Spark-TTS repo is required.",
next_step="Benchmark externally only with diacritized text/reference audio, then compare listenability and setup complexity against SILMA and Habibi.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Sofelia-TTS",
kind="tts",
role="Palestinian Arabic dialect voice benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/hamdallah/Sofelia-TTS",
hardware="GPU/large worker; MiraTTS runtime",
why="Apache-2.0 Palestinian Arabic TTS/voice-cloning model that may be useful for dialectal content, but it is not a standard MSA audiobook voice and its training data is private.",
next_step="Benchmark externally only if the target text is Palestinian/dialectal Arabic; keep SILMA/Habibi ahead for MSA books.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Arabic-F5-TTS-v2",
kind="tts",
role="personal/non-commercial MSA voice caution",
status="not default; non-commercial and requires tashkeel",
license="fair non-commercial research license",
source="https://huggingface.co/IbrahimSalah/Arabic-F5-TTS-v2",
hardware="GPU recommended; F5-TTS fine-tune",
why="Arabic MSA voice candidate with strong sample claims, but it requires fully diacritized Arabic and is not permissive enough for a default public website.",
next_step="Use only for personal listening tests unless the license is acceptable; do not wire as the default free permissive app voice.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="MOSS-TTS-Nano",
kind="tts",
role="CPU-friendly multilingual voice watchlist",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://github.com/OpenMOSS/MOSS-TTS-Nano",
hardware="CPU-friendly 0.1B model; ONNX path available",
why="Supports Arabic among 20 languages and has a lightweight ONNX CPU inference path, making it promising for a free worker if Arabic quality beats SILMA/Habibi.",
next_step="Export a cleaned Arabic sample, run the ONNX CLI with an Arabic reference voice, then compare listenability and runtime against SILMA on the same text.",
command="python scripts\\dry_run_pdf.py C:\\path\\to\\arabic-book.pdf --include-speech-text --speech-sample-chars 1200",
),
Candidate(
name="tts-arabic-onnx",
kind="tts",
role="compact Arabic-only ONNX voice benchmark",
status="not wired; benchmark externally first",
license="license not declared on model card/repo; confirm before production",
source="https://huggingface.co/nipponjo/tts-arabic-onnx",
hardware="CPU/ONNX-capable worker; small FastPitch/MixerTTS plus vocoder models",
why="Arabic-only ONNX package with FastPitch, tiny MixerTTS models, HiFi-GAN/Vocos vocoders, speaker and pace controls, and offline Python examples. It is promising for free CPU benchmarking, but license clarity must be resolved before production use.",
next_step="Export the same cleaned Arabic sample and benchmark FastPitch, MixerTTS, speaker IDs, pace, and vowelizer options against SILMA, Supertonic, MOSS-TTS-Nano, and Chatterbox ONNX before any app wiring.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Supertonic 3",
kind="tts",
role="wired optional fast CPU Arabic-capable voice",
status="wired optional sidecar",
license="OpenRAIL model, MIT sample code",
source="https://huggingface.co/Supertone/supertonic-3",
hardware="CPU-only ONNX; 99M model assets",
why="Supertonic 3 supports Arabic, runs locally with ONNX on CPU, and is much smaller than GPU-class multilingual voices, making it a practical free benchmark voice for long-book workers.",
next_step="Install the sidecar with scripts/setup_supertonic.ps1 or build with INSTALL_SUPERTONIC=1, then benchmark it against SILMA/Habibi on the same cleaned Arabic text.",
command="python scripts\\benchmark_voices.py --voices silma-local habibi-msa supertonic-ar espeak-ar-clear --text-file outputs\\external-tts-sample\\arabic-tts-sample.txt --out-dir outputs\\voice-benchmark --write-report outputs\\voice-benchmark-report.md",
),
Candidate(
name="Qwen3-TTS",
kind="tts",
role="strong open TTS to not promote for Arabic yet",
status="not Arabic-ready for this app",
license="Apache-2.0",
source="https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-Base",
hardware="GPU/large worker; 0.6B or 1.7B models",
why="Official Qwen3-TTS sources are strong and Apache-2.0, but the released model cards list 10 languages and do not include Arabic.",
next_step="Do not add it to the Arabic voice dropdown until an official Arabic-capable checkpoint or Arabic fine-tune is verified.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Saudi Arabic Qwen3-TTS",
kind="tts",
role="Saudi/KSA Arabic dialect voice benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/vadimbelsky/qwen3-TTS-KSA",
hardware="GPU/large worker; Qwen3-TTS 1.7B fine-tune",
why="Apache-2.0 Qwen3-TTS fine-tune for Saudi/KSA Arabic speech; useful when the target text should sound Gulf/Saudi rather than MSA audiobook narration.",
next_step="Benchmark only for Saudi/KSA dialect fit using the same cleaned Arabic text as SILMA/Habibi, then keep it external unless it wins listening tests and the worker can handle 1.7B-class runtime.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
Candidate(
name="Emirati Qwen3.5-TTS",
kind="tts",
role="Emirati/Gulf Qwen voice benchmark",
status="not wired; benchmark externally first",
license="Apache-2.0",
source="https://huggingface.co/vadimbelsky/qwen3.5-TTS-Emirati",
hardware="GPU/large worker; Qwen3-TTS fine-tune",
why="Apache-2.0 Qwen3-TTS-family Emirati Arabic fine-tune; a newer dialect benchmark beside the smaller Emirati VITS voice.",
next_step="Benchmark only when Gulf/Emirati pronunciation is desired; keep SILMA/Habibi first for MSA books and compare against Emirati VITS Male before wiring anything.",
command="python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf --env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample",
),
]
def candidates_by_kind(kind: str | None = None) -> list[Candidate]:
if not kind:
return list(CANDIDATES)
return [candidate for candidate in CANDIDATES if candidate.kind == kind]
def score_json_path(kind: str) -> str:
if kind == "ocr":
return "outputs\\external-ocr-sample\\external-ocr-score.json"
if kind == "preprocessor":
return "outputs\\external-tts-sample\\tts-preprocessor-score.json"
return "outputs\\external-tts-sample\\voice-listening-score.json"
def score_report_path(kind: str) -> str:
if kind == "ocr":
return "outputs\\external-ocr-sample\\external-ocr-score.md"
if kind == "preprocessor":
return "outputs\\external-tts-sample\\tts-preprocessor-score.md"
return "outputs\\external-tts-sample\\voice-listening-score.md"
def promotion_report_path(kind: str) -> str:
if kind == "ocr":
return "outputs\\external-ocr-sample\\model-promotion-gate.md"
if kind == "preprocessor":
return "outputs\\external-tts-sample\\preprocessor-promotion-gate.md"
return "outputs\\external-tts-sample\\voice-promotion-gate.md"
def gate_kind(candidate: Candidate) -> str:
role = candidate.role.lower()
if candidate.kind == "tts" and "preprocessor" in role:
return "preprocessor"
return candidate.kind
def command_with_score_json(command: str, kind: str) -> str:
lines: list[str] = []
for line in command.splitlines():
updated = line
if "score_external_ocr.py" in updated and "--write-json" not in updated:
updated += f" --write-json {score_json_path('ocr')}"
if "score_voice_listening.py" in updated and "--write-json" not in updated:
updated += f" --write-json {score_json_path('tts')}"
if "score_tts_preprocessor.py" in updated and "--write-json" not in updated:
updated += f" --write-json {score_json_path('preprocessor')}"
lines.append(updated)
return "\n".join(lines)
def voice_score_command() -> str:
return (
"python scripts\\score_voice_listening.py "
"--rating silma-local=5,4,4,5,5 --rating candidate=5,5,5,5,5 "
f"--write-report {score_report_path('tts')} --write-json {score_json_path('tts')}"
)
def preprocessor_score_command() -> str:
return (
"python scripts\\score_tts_preprocessor.py "
"--rating plain=4,5,5,4,4 --rating candidate=5,5,5,5,5 "
f"--write-report {score_report_path('preprocessor')} --write-json {score_json_path('preprocessor')}"
)
def candidate_slug(name: str) -> str:
slug = re.sub(r"[^a-z0-9]+", "-", name.lower()).strip("-")
return slug or "candidate"
def ocr_score_command(candidate: Candidate) -> str:
slug = candidate_slug(candidate.name)
return (
"python scripts\\score_external_ocr.py "
f"--candidate {slug}=outputs\\external-ocr-sample\\{slug}.txt "
"--baseline-json outputs\\external-ocr-sample\\wired-ocr-baseline.json "
f"--write-report {score_report_path('ocr')} --write-json {score_json_path('ocr')}"
)
def promotion_gate_command(candidate: Candidate) -> str:
kind = gate_kind(candidate)
return (
"python scripts\\model_promotion_gate.py "
f"--candidate-name \"{candidate.name}\" --kind {kind} --license \"{candidate.license}\" "
f"--score-json {score_json_path(kind)} "
"--same-sample --runtime-ok --privacy-ok --human-reviewed "
f"--write-report {promotion_report_path(kind)}"
)
def benchmark_command(candidate: Candidate) -> str:
command = command_with_score_json(candidate.command, gate_kind(candidate))
needs_gate = not is_production_default_safe(candidate) and "not Arabic-ready" not in candidate.status
if not needs_gate:
return command
if "model_promotion_gate.py" in command:
return command
if candidate.kind == "ocr" and "score_external_ocr.py" not in command:
command = f"{command}\n{ocr_score_command(candidate)}"
if candidate.kind == "tts" and "score_voice_listening.py" not in command and "score_tts_preprocessor.py" not in command:
if gate_kind(candidate) == "preprocessor":
command = f"{command}\n{preprocessor_score_command()}"
else:
command = f"{command}\n{voice_score_command()}"
return f"{command}\n{promotion_gate_command(candidate)}"
def build_report(candidates: list[Candidate]) -> str:
lines = [
"# Free Arabic OCR/TTS Research Watchlist",
"",
"Use this file to evaluate new free models without changing the production default first.",
"A candidate should become a default only after it wins on a representative Arabic book sample and produces listenable audio.",
"License policy: production/default entries must use a permissive or explicitly acceptable license; uncertain, non-commercial, Llama-licensed, and correction-risk entries stay benchmark-only until reviewed.",
"",
"| Candidate | Type | Role | Status | License | Hardware | Source |",
"| --- | --- | --- | --- | --- | --- | --- |",
]
for item in candidates:
lines.append(
f"| {item.name} | {item.kind} | {item.role} | {item.status} | "
f"{item.license} | {item.hardware} | {item.source} |"
)
lines.extend(["", "## Benchmark Steps", ""])
for item in candidates:
lines.extend(
[
f"### {item.name}",
"",
f"Why: {item.why}",
"",
f"Next step: {item.next_step}",
"",
"```powershell",
benchmark_command(item),
"```",
"",
]
)
return "\n".join(lines).rstrip() + "\n"
def names_by_status(candidates: list[Candidate], status: str) -> list[Candidate]:
return [candidate for candidate in candidates if candidate.status == status]
def benchmark_candidates(candidates: list[Candidate], kind: str) -> list[Candidate]:
return [
candidate
for candidate in candidates
if candidate.kind == kind and not is_production_default_safe(candidate) and "not Arabic-ready" not in candidate.status
]
def build_recommendation_report(candidates: list[Candidate]) -> str:
default_voice = next(candidate for candidate in candidates if candidate.name == "SILMA TTS")
optional_ocr = [candidate for candidate in names_by_status(candidates, "wired optional sidecar") if candidate.kind == "ocr"]
optional_tts = [candidate for candidate in names_by_status(candidates, "wired optional sidecar") if candidate.kind == "tts"]
ocr_benchmarks = benchmark_candidates(candidates, "ocr")[:12]
tts_benchmarks = benchmark_candidates(candidates, "tts")[:16]
lines = [
"# Recommended Free Arabic PDF To Audio Stack",
"",
"This is the compact decision report generated from the current research watchlist.",
"",
"## Use Now",
"",
"| Layer | Recommendation | Why |",
"| --- | --- | --- |",
"| Embedded PDFs | PyMuPDF text extraction first | It is free, fast, and avoids OCR errors when the PDF already contains usable Arabic text. |",
"| Scanned PDFs | `OCR_ENGINE=tesseract OCR_RENDER_ZOOM=2 TESSERACT_PSM=4` | It produced the most readable text on the 5-page Arabic OCR benchmark while staying much faster than the comparison modes. |",
f"| Default voice | {default_voice.name} | {default_voice.why} |",
"| Download/storage | Worker-local retained audio files | Free by default and avoids Vercel's 4.5 MB function payload limit; Hugging Face free CPU disk is 50 GB but non-persistent, so downloads are short-lived. |",
"| Hosted shape | Vercel shell plus Docker worker via `WORKER_BASE_URL` | Vercel serves the easy website while the worker handles large PDFs, OCR, and TTS on free CPU Space hardware when the job size is reasonable. |",
"",
"## Install First On A Stronger Worker",
"",
"| Candidate | Type | Why | Next Step |",
"| --- | --- | --- | --- |",
]
for candidate in optional_ocr + optional_tts:
lines.append(f"| {candidate.name} | {candidate.kind} | {candidate.why} | {candidate.next_step} |")
lines.extend(
[
"",
"## Benchmark Before Promoting",
"",
"These are promising free/open candidates, but they should not replace the default stack until they win on the same 5-page Arabic sample and same cleaned TTS text.",
"",
"### OCR",
"",
"| Candidate | License | Why It Stays Benchmark-Only |",
"| --- | --- | --- |",
]
)
for candidate in ocr_benchmarks:
lines.append(f"| {candidate.name} | {candidate.license} | {candidate.next_step} |")
lines.extend(["", "### TTS", "", "| Candidate | License | Why It Stays Benchmark-Only |", "| --- | --- | --- |"])
for candidate in tts_benchmarks:
lines.append(f"| {candidate.name} | {candidate.license} | {candidate.next_step} |")
lines.extend(
[
"",
"### Current Voice Priority",
"",
"Use SILMA first for the practical free Arabic audiobook voice. On a stronger worker, benchmark Habibi MSA and OmniVoice next. Keep KaniTTS benchmark-only until the `lfm1.0` Hugging Face license metadata is reconciled with the model-card Apache-2.0 text.",
]
)
lines.extend(
[
"",
"## Promotion Rule",
"",
"Promote a model only when all of these are true:",
"",
"1. It is free for the intended personal/family use.",
"2. Its license is acceptable for the deployment.",
"3. It beats the current stack on the same selected Arabic pages or same cleaned Arabic voice sample.",
"4. It preserves Arabic reading order, words, and pronunciation better than the default.",
"5. Its runtime is acceptable for the target worker.",
"6. The generated JSON score passes `scripts\\model_promotion_gate.py` after human review.",
"",
"Current practical default: PyMuPDF -> `tesseract@2x-psm4` OCR -> SILMA TTS -> downloadable worker audio.",
"",
]
)
return "\n".join(lines)
def build_decision_card(candidates: list[Candidate]) -> dict[str, object]:
optional_sidecars = [
candidate
for candidate in candidates
if candidate.status == "wired optional sidecar"
]
benchmark_ocr = benchmark_candidates(candidates, "ocr")
benchmark_tts = benchmark_candidates(candidates, "tts")
return {
"title": "Recommended Free Arabic PDF To Audio Decision Card",
"currentDefault": {
"pdf": "PyMuPDF embedded text first",
"scannedPdfOcr": "OCR_ENGINE=tesseract OCR_RENDER_ZOOM=2 TESSERACT_PSM=4",
"voice": "SILMA TTS (silma-local)",
"audioStorage": "worker-local retained downloads",
"hosting": "Vercel shell plus Docker worker via WORKER_BASE_URL",
},
"whyThisDefault": [
"Uses free/local software by default instead of paid cloud TTS.",
"Avoids OCR when embedded Arabic text already exists.",
"Uses Arabic-trained OCR sidecars when available and keeps Tesseract/PaddleOCR fallbacks.",
"Uses an Arabic-focused Apache-2.0 voice first, with eSpeak NG only as a fallback.",
"Keeps large PDF and audio work off Vercel Functions by delegating to the worker.",
],
"installFirstOnStrongerWorker": [
{
"name": candidate.name,
"kind": candidate.kind,
"license": candidate.license,
"why": candidate.why,
"nextStep": candidate.next_step,
}
for candidate in optional_sidecars
],
"benchmarkBeforePromoting": {
"ocrCount": len(benchmark_ocr),
"ttsCount": len(benchmark_tts),
"ocrFirst": [candidate.name for candidate in benchmark_ocr[:12]],
"ttsFirst": [candidate.name for candidate in benchmark_tts[:16]],
"voicePriority": "Use SILMA first, then benchmark Habibi MSA and OmniVoice on stronger hardware. Keep KaniTTS benchmark-only until its lfm1.0 Hugging Face license metadata is reconciled with the model-card Apache-2.0 text.",
},
"promotionGate": {
"sameSample": "same selected Arabic pages or same cleaned Arabic TTS text",
"scoreJsonRequired": True,
"tool": "scripts\\model_promotion_gate.py",
"requires": [
"candidate beats the current stack",
"license is acceptable for the deployment",
"worker runtime and memory are acceptable",
"privacy/API terms are clear",
"human Arabic review confirms faithful text/audio",
],
},
"commands": {
"refreshResearch": (
"python scripts\\refresh_research_evidence.py "
"--watchlist-report docs\\research-watchlist.md "
"--recommendation-report docs\\recommended-free-stack.md"
),
"sourceAudit": "python scripts\\check_research_sources.py --json",
"prepareBook": (
"python scripts\\prepare_book_workflow.py C:\\path\\to\\arabic-book.pdf "
"--sample-pages 1 --engine-preset balanced --verify-audio "
"--write-env outputs\\recommended-ocr.env --write-report outputs\\book-prep-report.md"
),
"externalOcrSample": (
"python scripts\\export_ocr_sample_images.py C:\\path\\to\\arabic-book.pdf "
"--out-dir outputs\\external-ocr-sample"
),
"externalTtsSample": (
"python scripts\\export_tts_sample.py C:\\path\\to\\arabic-book.pdf "
"--env-file outputs\\recommended-ocr.env --out-dir outputs\\external-tts-sample"
),
},
}
def build_decision_card_report(card: dict[str, object]) -> str:
current = card["currentDefault"] # type: ignore[index]
promotion = card["promotionGate"] # type: ignore[index]
benchmarks = card["benchmarkBeforePromoting"] # type: ignore[index]
commands = card["commands"] # type: ignore[index]
lines = [
"# Recommended Free Arabic PDF To Audio Decision Card",
"",
"## Current Default",
"",
"| Layer | Choice |",
"| --- | --- |",
]
for key, value in dict(current).items():
lines.append(f"| {key} | {value} |")
lines.extend(["", "## Why", ""])
for reason in card["whyThisDefault"]: # type: ignore[index]
lines.append(f"- {reason}")
lines.extend(
[
"",
"## Benchmark Queue",
"",
f"OCR benchmark candidates tracked: {dict(benchmarks)['ocrCount']}",
f"TTS benchmark candidates tracked: {dict(benchmarks)['ttsCount']}",
"",
"OCR first lane: " + ", ".join(dict(benchmarks)["ocrFirst"]), # type: ignore[arg-type]
"TTS first lane: " + ", ".join(dict(benchmarks)["ttsFirst"]), # type: ignore[arg-type]
"",
"## Promotion Gate",
"",
f"Tool: `{dict(promotion)['tool']}`",
f"Same sample: {dict(promotion)['sameSample']}",
"",
]
)
for item in dict(promotion)["requires"]: # type: ignore[index]
lines.append(f"- {item}")
lines.extend(["", "## Commands", ""])
for label, command in dict(commands).items():
lines.extend([f"{label}:", "", "```powershell", str(command), "```", ""])
return "\n".join(lines).rstrip() + "\n"
def write_report(path: Path, candidates: list[Candidate]) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(build_report(candidates), encoding="utf-8")
def write_recommendation_report(path: Path, candidates: list[Candidate]) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(build_recommendation_report(candidates), encoding="utf-8")
def write_decision_card_json(path: Path, candidates: list[Candidate]) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(json.dumps(build_decision_card(candidates), ensure_ascii=False, indent=2) + "\n", encoding="utf-8")
def write_decision_card_report(path: Path, candidates: list[Candidate]) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(build_decision_card_report(build_decision_card(candidates)), encoding="utf-8")
def main() -> None:
parser = argparse.ArgumentParser(description="Print or write the free Arabic OCR/TTS research watchlist.")
parser.add_argument("--kind", choices=["ocr", "tts"], help="Show only OCR or TTS candidates.")
parser.add_argument("--write-report", type=Path, default=DEFAULT_REPORT, help="Markdown report destination.")
parser.add_argument(
"--write-recommendation-report",
type=Path,
help="Write a compact recommended-stack decision report.",
)
parser.add_argument("--write-decision-card-json", type=Path, help="Write a machine-readable recommendation card.")
parser.add_argument("--write-decision-card-report", type=Path, help="Write a readable recommendation card.")
parser.add_argument("--json", action="store_true", help="Print JSON instead of Markdown.")
parser.add_argument("--check-license-policy", action="store_true", help="Fail if default-safe and benchmark-only license lanes are mixed.")
args = parser.parse_args()
selected = candidates_by_kind(args.kind)
violations = license_policy_violations(selected)
if args.check_license_policy and violations:
print(json.dumps({"ready": False, "violations": violations}, ensure_ascii=False, indent=2))
raise SystemExit(1)
if args.json:
print(
json.dumps(
{
"ready": not violations,
"violations": violations,
"candidates": [
{**asdict(candidate), "command": benchmark_command(candidate)} for candidate in selected
],
},
ensure_ascii=False,
indent=2,
)
)
return
write_report(args.write_report, selected)
if args.write_recommendation_report:
write_recommendation_report(args.write_recommendation_report, selected)
if args.write_decision_card_json:
write_decision_card_json(args.write_decision_card_json, selected)
if args.write_decision_card_report:
write_decision_card_report(args.write_decision_card_report, selected)
print(f"Wrote {args.write_report}")
if args.write_recommendation_report:
print(f"Wrote {args.write_recommendation_report}")
if args.write_decision_card_json:
print(f"Wrote {args.write_decision_card_json}")
if args.write_decision_card_report:
print(f"Wrote {args.write_decision_card_report}")
if __name__ == "__main__":
main()
|