Add language polish and seller shortcuts

README.md

@@ -19,13 +19,16 @@ The submission version is feature-frozen around the complete demo loop:
 
 - Record a transaction with your microphone and transcribe it with faster-whisper.
 - Configure seller context: business name, currency label, low-stock threshold, and language style.
+- Pick common currency presets for INR, USD, EUR, GBP, MXN, and BRL, or enter a custom currency label.
 - Use a first-run “Start in 60 seconds” guide for setup, first transaction, review, and closeout.
 - Type or paste a transaction note.
 - Bulk import multiple pasted notes for review and editing.
 - Parse it with Modal-hosted NVIDIA Nemotron when configured, with local rules as a deterministic fallback.
 - Parse common English, Hinglish, Hindi-lite, Gujarati-lite, Spanish, French, and Portuguese seller phrases with deterministic fallback rules.
 - Review a human-friendly transaction card with warning badges before saving.
+- See a language/confidence chip after parsing so multilingual notes are visibly handled.
 - Correct transaction type, item, customer, quantity, price, amount, notes, and confidence directly before saving.
+- Keep a correction log for review edits so model/rule mistakes become visible field-test evidence.
 - Choose `Cloud AI first` or `Local fallback only` from the Record screen to make the AI route explicit during demos.
 - Save the structured transaction to SQLite.
 - See a “Saved just now” receipt with bookkeeping side effects.
@@ -34,7 +37,8 @@ The submission version is feature-frozen around the complete demo loop:
 - Use a mobile-first, business-style Gradio interface with custom styling.
 - Run a Daily Closeout that prepares PDF, WhatsApp summary, and CSV exports together.
 - Download a Daily Summary PDF report.
-- Generate a WhatsApp-ready daily business summary.
+- Generate WhatsApp-ready daily business summaries in English, Spanish, French, or Portuguese.
+- Run shortcut commands like `close today`, `show Amit`, or `stock mangoes`.
 - Generate customer follow-up reminders and an inventory reorder list.
 - Capture field-test evidence with a seller checklist and anonymized feedback notes.
 - Offload speech transcription and LLM parsing to optional Modal endpoints.
@@ -106,7 +110,7 @@ The same examples are available in `sample_data/demo_transactions.txt`.
 2. Use the `Hackathon Demo Launchpad` on the first screen to seed demo transactions.
 3. Open `Record Text & Voice`, speak or type `Sold 12 mangoes, 20 each`, parse it, and review the human-readable transaction card.
 4. Correct one inline review field before saving, such as quantity or amount, to show the seller correction loop.
-5. Show the parse status and warning badges: Modal/NVIDIA Nemotron when available, local fallback when needed.
+5. Show the parse status, language/confidence chip, and warning badges: Modal/NVIDIA Nemotron when available, local fallback when needed.
 6. Save the transaction and show the Command Center plus the receipt with stock, customer, or amount side effects.
 7. Open `Dashboard`, `Customer Credit`, and `Inventory` to show the Insight Coach, timeline charts, customer detail, follow-up message, inventory detail, reorder list, and automatic bookkeeping updates.
 8. Open `Field Test` to show what a seller tried and what changed after feedback.
@@ -196,10 +200,14 @@ curl -X POST https://sagarpat3199--voiceledger-api.modal.run/parse \
 - Smart review warnings flag low confidence, missing fields, duplicate risk, and negative stock before save.
 - Inline review editing lets sellers correct the parsed transaction before the save touches the ledger.
 - Seller setup persists business name, currency label, low-stock threshold, and language style in SQLite.
+- Currency presets make the app easier to demo for INR, USD, EUR, GBP, MXN, BRL, and custom local labels.
 - Field Test persists anonymized seller evidence and a workflow checklist in SQLite.
+- The Field Test mistake log records corrected fields before save so product feedback is visible.
 - Local fallback rules include common English, Hinglish, Hindi-lite, Gujarati-lite, Spanish, French, and Portuguese seller notes.
+- Parse status includes a lightweight language/confidence chip for multilingual seller notes.
 - Customer follow-up and inventory reorder helpers generate WhatsApp-ready action messages.
-- WhatsApp summaries provide a short copyable daily recap for sharing.
+- WhatsApp summaries provide short copyable daily recaps in English, Spanish, French, or Portuguese.
+- Voice command shortcuts provide quick access to daily closeout, customer detail, and inventory detail.
 - Bulk import splits pasted notes by line, parses each line, supports review edits, and saves all reviewed transactions.
 - Modal integration lives in `backend/`; if endpoint URLs are not configured, local fallback stays active.
 - The Record screen includes a local-only AI mode that skips Modal while preserving the same transaction review and save flow.

docs/demo-video-script.md

@@ -16,7 +16,7 @@ On the first screen, show the `Judge Demo Flow` panel:
 
 Click `Seed Demo Transactions`.
 
-Show `Seller Setup` and mention that business name, currency, low-stock threshold, and language style are saved in SQLite.
+Show `Seller Setup` and mention that business name, currency preset, low-stock threshold, and language style are saved in SQLite.
 
 Show the first-run guide and multilingual examples so judges see English, Hinglish, Gujarati-lite, Spanish, French, and Portuguese seller notes.
 
@@ -50,7 +50,9 @@ Click parse. Show:
 - Inline review fields for correcting item, quantity, amount, customer, and notes before saving
 - Warning badges for low confidence, missing fields, duplicate risk, or negative stock when present
 - Parse source/status
+- Language/confidence chip for multilingual notes
 - Command Center update
+- Voice command shortcuts such as `close today`, `show Amit`, and `stock mangoes`
 - “Saved just now” receipt with stock, customer, or amount side effects
 
 ## 0:45-1:05 — Bookkeeping Updates
@@ -71,13 +73,13 @@ In `Ledger`, load a transaction by id, update or delete it, and show balances re
 
 Click `Download CSV`.
 
-Open `Reports & PDF`, run `Daily Closeout`, then generate the PDF, WhatsApp summary, and CSV export.
+Open `Reports & PDF`, run `Daily Closeout`, then generate the PDF, translated WhatsApp summary, and CSV export.
 
 Suggested screenshot/GIF moments:
 
 - Record flow: first-run guide, multilingual examples, review card, inline correction, parse source, warning badges, save receipt.
 - Seller setup: currency/threshold/language style and Command Center.
-- Field Test: seller checklist and feedback notes.
+- Field Test: seller checklist, feedback notes, and correction log.
 - Submission Story: AI pipeline and small-model fit card.
 - Dashboard: sales, expenses, profit, credit, timeline, top item, low-stock inventory.
 - Reports/Ledger: Daily Closeout, PDF download, WhatsApp summary, CSV export.

docs/field-notes.md

@@ -50,8 +50,9 @@ The hackathon version is intentionally frozen around the complete bookkeeping lo
 - Voice and text capture.
 - Modal/Nemotron parsing with local rule fallback.
 - Seller setup for business name, currency, low-stock threshold, and language style.
+- Currency presets and translated WhatsApp summaries for sellers who demo in INR, USD, EUR, GBP, MXN, BRL, English, Spanish, French, or Portuguese contexts.
 - Hinglish/Hindi-lite/Gujarati-lite/Spanish/French/Portuguese fallback examples for common seller notes.
-- First-run onboarding, multilingual examples, human-friendly review cards, inline review correction, warning badges, receipts, and save feedback.
+- First-run onboarding, multilingual examples, language/confidence chips, human-friendly review cards, inline review correction, mistake logging, warning badges, receipts, and save feedback.
 - Command Center, dashboard timeline, seller-day timeline, customer follow-up, inventory reorder list, ledger correction, Daily Closeout, PDF, CSV, and WhatsApp exports.
 - Field Test Mode for the seller checklist and anonymized “who/tried/changed” feedback evidence.
 - Demo Health for backend observability.
@@ -61,13 +62,13 @@ This freeze keeps the demo reliable and makes the value proposition easier to ju
 
 ## What We Learned
 
-The strongest product behavior is not the model call itself. It is the reliable loop around it: voice input, readable review, warning badges, save receipt, correction, and export. Informal sellers need speed, but they also need a way to spot mistakes and fix them. Edit/delete, detail drilldowns, Daily Closeout, and CSV export make the app credible as a bookkeeping tool instead of a parser demo.
+The strongest product behavior is not the model call itself. It is the reliable loop around it: voice input, readable review, warning badges, save receipt, correction, and export. Informal sellers need speed, but they also need a way to spot mistakes and fix them. The correction log makes those mistakes measurable during field tests, while edit/delete, detail drilldowns, Daily Closeout, and CSV export make the app credible as a bookkeeping tool instead of a parser demo.
 
 The small-model constraint helped keep the app honest. The model handles the fuzzy human input; the code owns the accounting state.
 
 ## Screenshot Moments
 
-- Record flow: first-run onboarding, multilingual examples, review card, inline corrections, warning badges, source/status, save receipt.
+- Record flow: first-run onboarding, multilingual examples, language/confidence chip, voice command shortcuts, review card, inline corrections, warning badges, source/status, save receipt.
 - Dashboard: daily totals, Insight Coach, outstanding credit, sales/expense timeline, top seller, low-stock inventory.
-- Field Test: workflow checklist and anonymized feedback evidence.
+- Field Test: workflow checklist, correction log, and anonymized feedback evidence.
 - Reports and exports: Daily Closeout, PDF download, WhatsApp summary, CSV ledger export.

docs/submission-checklist.md

@@ -17,13 +17,16 @@
 - `Demo Health` shows Modal reachable and backend version.
 - `Demo Health` shows NVIDIA Nemotron parser status.
 - Seller Setup saves business name, currency label, low-stock threshold, and language style.
+- Currency preset updates the currency label for INR, USD, EUR, GBP, MXN, and BRL.
 - Text parse works with `Sold 12 mangoes, 20 each`.
 - Local fallback parses `Amit ne 100 dene hai`, `Amit ne 50 diya`, and `50 mango kharida`.
 - Voice parse works and transcript appears.
 - Review card shows type, item, quantity, price, amount, customer, source, and confidence.
+- Parse status shows source plus language/confidence.
 - Smart warnings appear for low confidence, missing fields, duplicate risk, and negative stock.
 - Save shows a receipt with transaction type, amount, and side effects.
 - Command Center updates after save and seed.
+- Voice command shortcuts work for `close today`, `show Amit`, and `stock mangoes`.
 - Save updates Dashboard and Ledger.
 - Dashboard timeline loads from saved transactions.
 - Seller-day timeline shows recent saved transactions.
@@ -37,7 +40,8 @@
 - CSV export downloads from Ledger.
 - Daily Closeout generates PDF, CSV, WhatsApp summary, and a status line.
 - PDF report downloads from Reports.
-- WhatsApp summary generates copyable text.
+- WhatsApp summary generates copyable text in English, Spanish, French, and Portuguese.
+- Field Test correction log shows edited review fields.
 
 ## Prize Alignment
 

tests/test_ui_audio_flow.py

@@ -97,6 +97,23 @@ def test_parse_note_local_mode_surfaces_local_fallback(monkeypatch) -> None:
     assert "Safe to save" in review_card
 
 
+def test_parse_note_surfaces_language_confidence_chip(monkeypatch) -> None:
+    monkeypatch.setattr(
+        gradio_app.modal_api,
+        "parse_transaction_result",
+        lambda text, fallback, **kwargs: ParseResult(
+            transaction=fallback(text),
+            source="local",
+            message="Parsed locally with the rule parser.",
+        ),
+    )
+
+    _, _, status, _ = gradio_app._parse_note("Vendí 12 mangos, 20 cada uno")
+
+    assert "Spanish" in status
+    assert "High confidence" in status
+
+
 def test_high_contrast_demo_panels_are_rendered() -> None:
     panel = gradio_app._info_panel(
         "Demo Health",
@@ -210,6 +227,9 @@ def test_apply_review_edits_updates_payload_and_review_card(tmp_path) -> None:
     assert state == payload
     assert "Review updated" in status
     assert "250" in review_card
+    correction_log = gradio_app.get_correction_log(db_path)
+    assert len(correction_log) == 1
+    assert "quantity" in correction_log.iloc[0]["changed_fields"]
 
 
 def test_receipt_card_summarizes_saved_sale(tmp_path) -> None:
@@ -267,6 +287,18 @@ def test_customer_followup_and_reorder_helpers(tmp_path) -> None:
     assert "Onions" in message
 
 
+def test_currency_presets_and_voice_commands(tmp_path) -> None:
+    db_path = tmp_path / "voiceledger.sqlite3"
+    gradio_app.add_transaction(gradio_app.local_parse_transaction("Amit owes 100"), db_path)
+    gradio_app.add_transaction(gradio_app.local_parse_transaction("Bought 3 onions"), db_path)
+
+    assert gradio_app._currency_symbol_for_preset("Brazil - BRL (R$)") == "R$"
+    assert gradio_app._currency_preset_for_symbol("€") == "European Union - EUR (€)"
+    assert "Amit" in gradio_app._run_voice_command("show Amit", db_path)
+    assert "onions" in gradio_app._run_voice_command("stock onions", db_path)
+    assert "Daily Closeout Ready" in gradio_app._run_voice_command("close today", db_path)
+
+
 def test_insight_coach_surfaces_credit_and_stock_actions(tmp_path) -> None:
     db_path = tmp_path / "voiceledger.sqlite3"
     gradio_app.add_transaction(gradio_app.local_parse_transaction("Amit owes 100"), db_path)

tests/test_whatsapp_summary.py

@@ -35,3 +35,19 @@ def test_generate_whatsapp_summary_handles_empty_data(tmp_path: Path) -> None:
     assert "Outstanding Credit: ₹0" in summary
     assert "Top Product: None" in summary
     assert "Low Stock: None" in summary
+
+
+def test_generate_whatsapp_summary_supports_language_labels(tmp_path: Path) -> None:
+    db_path = tmp_path / "voiceledger.sqlite3"
+    add_transaction(parse_transaction("Vendí 12 mangos, 20 cada uno"), db_path)
+
+    spanish = generate_whatsapp_summary(db_path=db_path, language="Spanish", currency_symbol="$")
+    french = generate_whatsapp_summary(db_path=db_path, language="French", currency_symbol="€")
+    portuguese = generate_whatsapp_summary(db_path=db_path, language="Portuguese", currency_symbol="R$")
+
+    assert "Resumen Diario" in spanish
+    assert "Ventas: $240" in spanish
+    assert "Resume Quotidien" in french
+    assert "Ventes: €240" in french
+    assert "Resumo Diario" in portuguese
+    assert "Vendas: R$240" in portuguese

voiceledger/ledger/corrections.py

@@ -0,0 +1,107 @@
+"""Correction log persistence for parsed transaction review edits."""
+
+from __future__ import annotations
+
+import json
+import sqlite3
+from datetime import datetime
+from pathlib import Path
+from typing import Any
+
+import pandas as pd
+
+from voiceledger.config import get_database_path
+
+
+CORRECTION_SCHEMA_SQL = """
+CREATE TABLE IF NOT EXISTS correction_log (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    original_payload TEXT NOT NULL,
+    corrected_payload TEXT NOT NULL,
+    changed_fields TEXT NOT NULL,
+    created_at TEXT NOT NULL DEFAULT CURRENT_TIMESTAMP
+);
+"""
+
+CORRECTION_COLUMNS = ["id", "changed_fields", "original_payload", "corrected_payload", "created_at"]
+
+
+def initialize_correction_log_table(db_path: str | Path | None = None) -> Path:
+    """Create the correction log table if needed."""
+    path = _resolve_db_path(db_path)
+    path.parent.mkdir(parents=True, exist_ok=True)
+    with sqlite3.connect(path) as connection:
+        connection.execute(CORRECTION_SCHEMA_SQL)
+        connection.commit()
+    return path
+
+
+def record_correction(
+    original_payload: dict[str, Any],
+    corrected_payload: dict[str, Any],
+    db_path: str | Path | None = None,
+) -> int | None:
+    """Persist a correction when review edits changed transaction fields."""
+    changed_fields = _changed_fields(original_payload, corrected_payload)
+    if not changed_fields:
+        return None
+
+    path = initialize_correction_log_table(db_path)
+    with sqlite3.connect(path) as connection:
+        cursor = connection.execute(
+            """
+            INSERT INTO correction_log (
+                original_payload,
+                corrected_payload,
+                changed_fields,
+                created_at
+            )
+            VALUES (?, ?, ?, ?)
+            """,
+            (
+                json.dumps(original_payload, sort_keys=True),
+                json.dumps(corrected_payload, sort_keys=True),
+                ", ".join(changed_fields),
+                datetime.now().isoformat(sep=" ", timespec="seconds"),
+            ),
+        )
+        connection.commit()
+        return int(cursor.lastrowid)
+
+
+def get_correction_log(db_path: str | Path | None = None) -> pd.DataFrame:
+    """Return saved parse correction rows."""
+    path = initialize_correction_log_table(db_path)
+    with sqlite3.connect(path) as connection:
+        rows = connection.execute(
+            """
+            SELECT id, changed_fields, original_payload, corrected_payload, created_at
+            FROM correction_log
+            ORDER BY id DESC
+            """
+        ).fetchall()
+    records = [dict(zip(CORRECTION_COLUMNS, row, strict=True)) for row in rows]
+    return pd.DataFrame.from_records(records, columns=CORRECTION_COLUMNS)
+
+
+def _changed_fields(original_payload: dict[str, Any], corrected_payload: dict[str, Any]) -> list[str]:
+    """Return fields whose values changed after review editing."""
+    tracked_fields = (
+        "transaction_type",
+        "item",
+        "quantity",
+        "unit_price",
+        "amount",
+        "customer",
+        "payment_status",
+        "notes",
+        "confidence",
+    )
+    return [field for field in tracked_fields if original_payload.get(field) != corrected_payload.get(field)]
+
+
+def _resolve_db_path(db_path: str | Path | None) -> Path:
+    """Resolve an explicit or configured database path."""
+    if db_path is None:
+        return get_database_path()
+    return Path(db_path).expanduser()

voiceledger/ledger/database.py

@@ -11,6 +11,7 @@ from typing import Any
 import pandas as pd
 
 from voiceledger.config import get_database_path
+from voiceledger.ledger.corrections import initialize_correction_log_table
 from voiceledger.ledger.customers import add_credit, initialize_customers_table, record_payment
 from voiceledger.ledger.inventory import add_stock, initialize_inventory_table, remove_stock
 from voiceledger.ledger.settings import initialize_business_settings_table
@@ -60,6 +61,7 @@ def initialize_database(db_path: str | Path | None = None) -> Path:
     initialize_customers_table(path)
     initialize_inventory_table(path)
     initialize_business_settings_table(path)
+    initialize_correction_log_table(path)
     return path
 
 

voiceledger/reports/whatsapp_summary.py

@@ -23,8 +23,10 @@ def generate_whatsapp_summary(
     low_stock_threshold: float = 5,
     business_name: str = "VoiceLedger",
     currency_symbol: str = "₹",
+    language: str = "English",
 ) -> str:
     """Generate a concise daily summary suitable for WhatsApp sharing."""
+    labels = _summary_labels(language)
     sales = calculate_daily_sales(db_path=db_path, report_date=report_date)
     expenses = calculate_daily_expenses(db_path=db_path, report_date=report_date)
     profit = calculate_net_profit(db_path=db_path, report_date=report_date)
@@ -34,16 +36,16 @@ def generate_whatsapp_summary(
 
     return "\n".join(
         [
-            f"{business_name} Daily Summary",
+            f"{business_name} {labels['daily_summary']}",
             "",
-            f"Sales: {_format_money(sales, currency_symbol)}",
-            f"Expenses: {_format_money(expenses, currency_symbol)}",
-            f"Profit: {_format_money(profit, currency_symbol)}",
+            f"{labels['sales']}: {_format_money(sales, currency_symbol)}",
+            f"{labels['expenses']}: {_format_money(expenses, currency_symbol)}",
+            f"{labels['profit']}: {_format_money(profit, currency_symbol)}",
             "",
-            f"Outstanding Credit: {_format_money(credit, currency_symbol)}",
+            f"{labels['outstanding_credit']}: {_format_money(credit, currency_symbol)}",
             "",
-            f"Top Product: {top_product}",
-            f"Low Stock: {low_stock}",
+            f"{labels['top_product']}: {top_product}",
+            f"{labels['low_stock']}: {low_stock}",
         ]
     )
 
@@ -68,3 +70,47 @@ def _format_money(value: float, currency_symbol: str) -> str:
     if amount.is_integer():
         return f"{currency_symbol}{int(amount)}"
     return f"{currency_symbol}{amount:,.2f}"
+
+
+def _summary_labels(language: str) -> dict[str, str]:
+    """Return localized labels for a WhatsApp daily summary."""
+    normalized = (language or "English").strip().lower()
+    labels = {
+        "english": {
+            "daily_summary": "Daily Summary",
+            "sales": "Sales",
+            "expenses": "Expenses",
+            "profit": "Profit",
+            "outstanding_credit": "Outstanding Credit",
+            "top_product": "Top Product",
+            "low_stock": "Low Stock",
+        },
+        "spanish": {
+            "daily_summary": "Resumen Diario",
+            "sales": "Ventas",
+            "expenses": "Gastos",
+            "profit": "Ganancia",
+            "outstanding_credit": "Credito Pendiente",
+            "top_product": "Producto Principal",
+            "low_stock": "Bajo Stock",
+        },
+        "french": {
+            "daily_summary": "Resume Quotidien",
+            "sales": "Ventes",
+            "expenses": "Depenses",
+            "profit": "Profit",
+            "outstanding_credit": "Credit En Attente",
+            "top_product": "Meilleur Produit",
+            "low_stock": "Stock Bas",
+        },
+        "portuguese": {
+            "daily_summary": "Resumo Diario",
+            "sales": "Vendas",
+            "expenses": "Despesas",
+            "profit": "Lucro",
+            "outstanding_credit": "Credito Pendente",
+            "top_product": "Produto Principal",
+            "low_stock": "Estoque Baixo",
+        },
+    }
+    return labels.get(normalized, labels["english"])

voiceledger/ui/gradio_app.py

@@ -28,6 +28,7 @@ from voiceledger.ledger.analytics import (
     top_selling_items,
 )
 from voiceledger.ledger.customers import get_customer_balances
+from voiceledger.ledger.corrections import get_correction_log, record_correction
 from voiceledger.ledger.database import (
     add_transaction,
     delete_transaction,
@@ -68,6 +69,16 @@ LANGUAGE_STYLE_CHOICES = [
     "Portuguese",
     "Multilingual",
 ]
+WHATSAPP_LANGUAGE_CHOICES = ["English", "Spanish", "French", "Portuguese"]
+CURRENCY_PRESETS = {
+    "India - INR (₹)": "₹",
+    "United States - USD ($)": "$",
+    "European Union - EUR (€)": "€",
+    "United Kingdom - GBP (£)": "£",
+    "Mexico - MXN ($)": "$",
+    "Brazil - BRL (R$)": "R$",
+    "Custom": "",
+}
 DEMO_NOTES = [
     "Bought 60 mangoes",
     "Sold 12 mangoes, 20 each",
@@ -128,11 +139,22 @@ def create_app(db_path: str | Path | None = None) -> gr.Blocks:
                         value=initial_settings["business_name"],
                         elem_classes="vl-panel",
                     )
+                    currency_preset_input = gr.Dropdown(
+                        choices=list(CURRENCY_PRESETS.keys()),
+                        label="Currency preset",
+                        value=_currency_preset_for_symbol(initial_settings["currency_symbol"]),
+                        elem_classes="vl-panel",
+                    )
                     currency_symbol_input = gr.Textbox(
                         label="Currency label",
                         value=initial_settings["currency_symbol"],
                         elem_classes="vl-panel",
                     )
+                currency_preset_input.change(
+                    fn=_currency_symbol_for_preset,
+                    inputs=currency_preset_input,
+                    outputs=currency_symbol_input,
+                )
                 with gr.Row():
                     low_stock_threshold_input = gr.Number(
                         label="Low-stock threshold",
@@ -186,6 +208,15 @@ def create_app(db_path: str | Path | None = None) -> gr.Blocks:
                 gr.HTML(
                     _multilingual_examples_panel()
                 )
+                gr.HTML(_section_heading("Voice Command Shortcuts"))
+                with gr.Row():
+                    command_input = gr.Textbox(
+                        label="Command",
+                        placeholder="close today, show Amit, stock mangoes",
+                        elem_classes="vl-panel",
+                    )
+                    command_button = gr.Button("Run Command")
+                command_output = gr.HTML(_empty_detail_card("Command result", "Try close today, show Amit, or stock mangoes."))
                 with gr.Row(elem_classes="vl-example-row"):
                     example_sale_button = gr.Button("Try sale")
                     example_expense_button = gr.Button("Try expense")
@@ -298,6 +329,11 @@ def create_app(db_path: str | Path | None = None) -> gr.Blocks:
                     inputs=None,
                     outputs=[note_input, record_demo_status],
                 )
+                command_button.click(
+                    fn=lambda command: _run_voice_command(command, db_path),
+                    inputs=command_input,
+                    outputs=command_output,
+                )
                 apply_review_edits_button.click(
                     fn=lambda parsed, transaction_type, item, quantity, unit_price, amount, customer, payment_status, notes, confidence: _apply_review_edits(
                         parsed,
@@ -505,6 +541,21 @@ def create_app(db_path: str | Path | None = None) -> gr.Blocks:
                     value=_field_test_summary(initial_settings),
                     elem_classes="vl-status",
                 )
+                gr.HTML(_section_heading("Mistake Log"))
+                refresh_corrections_button = gr.Button("Refresh Correction Log")
+                correction_log_output = gr.Dataframe(
+                    value=get_correction_log(db_path),
+                    headers=["id", "changed_fields", "original_payload", "corrected_payload", "created_at"],
+                    label="Corrections made before save",
+                    interactive=False,
+                    wrap=True,
+                    elem_classes="vl-panel",
+                )
+                refresh_corrections_button.click(
+                    fn=lambda: get_correction_log(db_path),
+                    inputs=None,
+                    outputs=correction_log_output,
+                )
 
         with gr.Column(visible=False, elem_classes="vl-page-section") as bulk_page:
                 gr.HTML('<div id="vl-page-bulk" class="vl-page-anchor"></div>')
@@ -700,6 +751,12 @@ def create_app(db_path: str | Path | None = None) -> gr.Blocks:
                 )
                 gr.HTML(_section_heading("WhatsApp Summary"))
                 generate_whatsapp_button = gr.Button("Generate WhatsApp Summary")
+                whatsapp_language_input = gr.Dropdown(
+                    choices=WHATSAPP_LANGUAGE_CHOICES,
+                    label="Summary language",
+                    value="English",
+                    elem_classes="vl-panel",
+                )
                 whatsapp_summary_output = gr.Textbox(
                     label="WhatsApp Summary",
                     lines=10,
@@ -717,8 +774,8 @@ def create_app(db_path: str | Path | None = None) -> gr.Blocks:
                     """
                 )
                 generate_whatsapp_button.click(
-                    fn=lambda: _generate_whatsapp_summary(db_path),
-                    inputs=None,
+                    fn=lambda language: _generate_whatsapp_summary(db_path, language),
+                    inputs=whatsapp_language_input,
                     outputs=whatsapp_summary_output,
                 )
 
@@ -1249,6 +1306,20 @@ def _ai_mode_status(ai_mode: str | None) -> str:
     """
 
 
+def _currency_symbol_for_preset(preset: str | None) -> str:
+    """Return the currency symbol for a selected preset."""
+    return CURRENCY_PRESETS.get(preset or "Custom", "")
+
+
+def _currency_preset_for_symbol(symbol: str | None) -> str:
+    """Return a likely currency preset for an existing symbol."""
+    cleaned = (symbol or "").strip()
+    for preset, preset_symbol in CURRENCY_PRESETS.items():
+        if preset != "Custom" and preset_symbol == cleaned:
+            return preset
+    return "Custom"
+
+
 def _demo_health_placeholder() -> pd.DataFrame:
     """Return placeholder health rows so the demo health table is never blank."""
     return pd.DataFrame(
@@ -1395,6 +1466,7 @@ def _apply_review_edits(
         confidence=confidence,
     )
     payload = transaction.model_dump()
+    record_correction(transaction_payload, payload, db_path)
     warnings = _review_warnings(transaction, db_path)
     status = _status_message(transaction, "Review updated from your edits.", warnings=warnings)
     return payload, payload, status, _review_card(transaction, "Review updated from your edits.", warnings)
@@ -1946,6 +2018,36 @@ def _get_inventory_detail(item: str | None, db_path: str | Path | None) -> tuple
     return summary, matches[columns].reset_index(drop=True)
 
 
+def _run_voice_command(command: str | None, db_path: str | Path | None) -> str:
+    """Run a lightweight text/voice command shortcut."""
+    cleaned = " ".join(str(command or "").split()).strip()
+    if not cleaned:
+        return _empty_detail_card("Command result", "Try close today, show Amit, or stock mangoes.")
+
+    lowered = cleaned.lower()
+    if lowered in {"close today", "daily closeout", "closeout", "summary"}:
+        summary, pdf_path, csv_path, whatsapp, status = _run_daily_closeout(db_path)
+        file_note = f"PDF: {pdf_path or 'needs attention'}. CSV: {csv_path or 'needs attention'}."
+        return f"{summary}<section class=\"vl-detail-card\"><h2>Command result</h2><p>{escape(status)} {escape(file_note)}</p><pre>{escape(whatsapp)}</pre></section>"
+
+    if lowered.startswith(("show ", "customer ")):
+        customer = cleaned.split(" ", 1)[1] if " " in cleaned else ""
+        summary, _ = _get_customer_detail(customer, db_path)
+        return summary
+
+    if lowered.startswith(("stock ", "inventory ")):
+        item = cleaned.split(" ", 1)[1] if " " in cleaned else ""
+        summary, _ = _get_inventory_detail(item, db_path)
+        return summary
+
+    parsed = local_parse_transaction(cleaned)
+    if parsed.transaction_type != "unknown":
+        warnings = _review_warnings(parsed, db_path)
+        return _review_card(parsed, "Command parsed as a transaction. Copy it into the transaction note to save.", warnings)
+
+    return _empty_detail_card("Command result", "Command not recognized. Try close today, show Amit, stock mangoes, or a transaction note.")
+
+
 def _generate_daily_summary_report(db_path: str | Path | None) -> tuple[str | None, str]:
     """Generate the Daily Summary PDF for download in Gradio."""
     settings = get_business_settings(db_path)
@@ -2093,7 +2195,7 @@ def _metric_card(label: str, value: str, note: str, profit: float | None = None)
     """
 
 
-def _generate_whatsapp_summary(db_path: str | Path | None) -> str:
+def _generate_whatsapp_summary(db_path: str | Path | None, language: str = "English") -> str:
     """Generate a WhatsApp summary using seller settings."""
     settings = get_business_settings(db_path)
     return generate_whatsapp_summary(
@@ -2101,6 +2203,7 @@ def _generate_whatsapp_summary(db_path: str | Path | None) -> str:
         low_stock_threshold=get_low_stock_threshold(db_path),
         business_name=settings["business_name"],
         currency_symbol=settings["currency_symbol"],
+        language=language,
     )
 
 
@@ -2327,6 +2430,7 @@ def _status_message(
     """Return a human-readable parsing status."""
     parts = []
     parts.append(_source_chip(source_message, fallback_reason))
+    parts.append(_language_confidence_chip(transaction))
     if prefix:
         parts.append(prefix)
     parts.append(source_message)
@@ -2351,6 +2455,29 @@ def _source_chip(source_message: str, fallback_reason: str | None = None) -> str
     return '<span class="vl-status-chip">Parsed</span>'
 
 
+def _language_confidence_chip(transaction: Transaction) -> str:
+    """Return a compact language/confidence chip for parse status."""
+    language = _detect_note_language(transaction.notes)
+    confidence_label = "High confidence" if transaction.confidence >= 0.85 else "Needs review"
+    return f'<span class="vl-status-chip vl-status-chip-language">{escape(language)} · {confidence_label}</span>'
+
+
+def _detect_note_language(note: str | None) -> str:
+    """Detect a lightweight language label from common seller note phrases."""
+    text = (note or "").lower()
+    if any(token in text for token in ("vendí", "pagué", " debe ", " pagó", " cada uno", "suministros")):
+        return "Spanish"
+    if any(token in text for token in ("vendu", "payé", " doit ", " a payé", "chacun", "fournitures", "acheté")):
+        return "French"
+    if any(token in text for token in ("vendi", "paguei", " deve ", "pagou", "suprimentos", "comprei")):
+        return "Portuguese"
+    if any(token in text for token in ("dene hai", "dena hai", "diya", "kharida", "chukaya")):
+        return "Hinglish"
+    if any(token in text for token in ("lidha", "aapva che", "apvana che", "aapya")):
+        return "Gujarati-lite"
+    return "English"
+
+
 def _transcription_status(result: modal_api.TranscriptionResult) -> str:
     """Return a concise transcription source status."""
     status = result.message

voiceledger/ui/theme.py

@@ -637,6 +637,23 @@ body,
   color: #5f370e !important;
 }
 
+.vl-status-chip-language {
+  background: #f0eaff !important;
+  border-color: #bda8ef !important;
+  color: #43236d !important;
+}
+
+.vl-detail-card pre {
+  background: #f6f7f2 !important;
+  border: 1px solid var(--vl-border) !important;
+  border-radius: 10px !important;
+  color: var(--vl-text) !important;
+  font-size: 13px !important;
+  overflow-x: auto !important;
+  padding: 10px !important;
+  white-space: pre-wrap !important;
+}
+
 .vl-chip {
   background: #2c2926 !important;
   border-radius: 6px !important;