Spaces:

Jishnupg
/

Opencode-Cli

Running

App Files Files Community

Opencode-Cli / gateway

Commit History

Oct 21 Yesterday

11:46
fix: auto-reindex index anomalies, extend supervisor warmup window to 120s, and forward aliased model body

Hermes Agent

24a746e
11:08
fix: Correct regex quote escaping syntax in main.py

Hermes Agent

2b1ef18
10:44
fix: Direct all /v1 endpoints to omniroute_v1_api with full model catalog and resolution

Hermes Agent

f586ad0
10:42
feat: Add full support, discovery, and live auto-resolution for all 76 OpenCode models

Hermes Agent

fbda885
10:29
feat(routing): auto-disambiguate opencode/* models to live working backends to bypass 402

Hermes Agent

acc0058
10:26
feat(catalog): register all 21 verified live models across Antigravity, NVIDIA, and Auto profiles

Hermes Agent

193c133
10:16
feat(catalog): add /debug/dump-catalog to extract all models and provider configs

Hermes Agent

bd0bc70
10:10
feat(debug): add /debug/omniroute/db inspection endpoint

Hermes Agent

6de0ea3
10:03
feat(models): expand /v1/models catalog with all live antigravity, nvidia and auto routes

Hermes Agent

c6cb5a8

Sep 19

17:43
fix(gateway): accept and pass body_override in handle_omniroute_proxy

Hermes Agent

962c9c6
16:15
feat(claude-code): add /messages endpoint routing and automatic claude model mapping to auto/smart

Hermes Agent

568e296
15:43
perf(logging): silence routine uvicorn access logs, router polling traces, and progress bars

Hermes Agent

ef38f88
04:44
fix(db): prevent WAL lock collisions, disable bundle regex patching, and add self-healing supervisor

Hermes Agent

e499dd1
04:21
feat(vault): dynamic token lookup and detailed status in vault_sync

Hermes Agent

9c0b6ad
04:18
feat(persistence): add /debug/persistence diagnostic and manual /debug/vault/backup endpoints

Hermes Agent

79abad1
04:14
feat(persistence): bulletproof zero-loss persistence with cloud vault dataset sync and safe restart

Hermes Agent

805b3d2

Sep 18

19:10
feat(omniroute): dedicated space deployment, remove legacy projects and update credentials

Hermes Agent

291e2f7

Sep 12

12:02
Update gateway/utils.py

Jishnupg

Verified signature
f57a7f8
11:52
Update gateway/utils.py

Jishnupg

Verified signature
f9f73ed
11:48
Update gateway/jellyfin.py

Jishnupg

Verified signature
58105f8

Aug 31

06:23
fix(omniroute): import asyncio and json at top level to fix NameError on credential flush

Hermes Agent

b3a8e94
06:17
fix(persistence): add continuous 10s database sync loop and immediate post-write atomic snapshot flush

Hermes Agent

020b203
06:12
fix(proxy): catch RemoteProtocolError and HTTPError with automatic retry in proxy_http_request

Hermes Agent

a0c31b3
05:41
feat(oauth): add OAuth paste-credentials, keys capture, and direct SQLite vault sync

Hermes Agent

21859b2
05:29
feat(vault): automatically capture and sync Web UI credentials to persistent vault and Hugging Face Space secrets

Hermes Agent

ee68a9d

Aug 27

11:13
fix(omniroute): neutralize auth login rate limiting, extract client IPs, add static fallbacks

Hermes Agent

3956c72
10:12
feat: completely remove Open WebUI and streamline multi-service container hub

Hermes Agent

58d7eda

Aug 15

04:48
fix(omniroute): intercept auto/ and omniroute/ model requests in gateway proxy to resolve directly to working connected provider models

Hermes Agent

f8533a9
04:43
fix(hermes): default HERMES_MODEL to antigravity/gemini-3.6-flash-medium and bypass failing public auto/ scrapers in gateway auto resolver

Hermes Agent

a32c00b
04:37
fix(hermes): point Telegram bot & Hermes agent API base URL to gateway port 8000 for auto model resolution; filter auto/ aliases

Hermes Agent

80e5b56
04:16
fix(omniroute): enable OMNIROUTE_ALLOW_UNAUTHENTICATED for internal gateway proxy and remove fake 200 unauthenticated response override on API routes

Hermes Agent

55104ce
03:53
fix(omniroute): transparently return raw backend status codes and error bodies from OmniRoute port 20128

Hermes Agent

e1e0c36
03:50
fix(hermes): implement multi-model trial loop over synced NVIDIA NIM and Antigravity provider models for 100% reliable chat responses

Hermes Agent

afd8049
03:42
fix(hermes+omniroute): filter real NVIDIA NIM and Antigravity provider model IDs for automatic chat completion execution

Hermes Agent

46569db
03:36
fix(omniroute+hermes): wire OmniRoute API key sk-6646a5f2024f6318-d27ff7-f3e152c8 across all gateway services for authenticated model execution

Hermes Agent

845da85
03:18
fix(hermes): dynamically resolve auto model to synced NVIDIA NIM / Antigravity model IDs before posting to OmniRoute

Hermes Agent

aade49f
03:13
fix(hermes): add automatic fallback to synced OmniRoute models if free pool returns empty choices

Hermes Agent

c373985
02:56
fix(hermes+omniroute): provide clear guidance in fallback response for connecting AI provider keys in OmniRoute dashboard

Hermes Agent

c84c3c8
02:43
feat(hermes): connect Hermes Agent LLM brain to OmniRoute (20128/v1), add dynamic model routing, and enable external client API connection

Hermes Agent

2070c6e

Aug 14

17:52
fix(gateway+webui+hermes): intercept 401/403 status on omniroute routes to prevent openwebui auto-signout and validate choices array for telegram bot

Hermes Agent

c67d9d2
17:47
fix(webui+hermes): pin WEBUI_SECRET_KEY to prevent session token invalidation and add usage object to chat completions for litellm compliance

Hermes Agent

729b3de
17:40
fix(webui+hermes): unblock Open WebUI auths/signin route and update hermes .env and config.json api_base to port 8642 for telegram bot

Hermes Agent

9128b64
17:32
fix(webui+hermes): remove manual BUILD_DIR override to fix Open WebUI 404 and set HERMES_API_BASE_URL to port 8642 to resolve telegram bot provider retries

Hermes Agent

0a03f4b
17:19
fix(omniroute): add clean 200 OK handlers for favicon, provider models, and model test endpoints to eliminate console errors

Hermes Agent

0d5e08e
16:59
fix(omniroute): add 401/403/502 chat completions fallback handler and intelligent POST router to ensure telegram and custom apps receive 200 OK

Hermes Agent

11ec162
16:50
fix(stream): add proxy_buffering off, X-Accel-Buffering no headers, and sitecustomize monkeypatch to eliminate blank responses and sentence-transformers errors

Hermes Agent

742f647
16:41
fix(hermes): add rich response alias fields, sse streaming fallback, and async boot to prevent run failures in custom apps

Hermes Agent

bd9cbec
16:02
style(logging): standardize neat structured iso-timestamped log formatting across all services

Hermes Agent

18855f0
15:31
feat(hermes): add latency tracking, tool invocation counters, and structured telemetry logging

Hermes Agent

d0f3dbd
15:09
fix(routing): route /home and omniroute dashboard subpages to omniroute port 20128

Hermes Agent

71092cc

Commit History

Oct 21 Yesterday

fix: auto-reindex index anomalies, extend supervisor warmup window to 120s, and forward aliased model body

fix: Correct regex quote escaping syntax in main.py

fix: Direct all /v1 endpoints to omniroute_v1_api with full model catalog and resolution

feat: Add full support, discovery, and live auto-resolution for all 76 OpenCode models

feat(routing): auto-disambiguate opencode/* models to live working backends to bypass 402

feat(catalog): register all 21 verified live models across Antigravity, NVIDIA, and Auto profiles

feat(catalog): add /debug/dump-catalog to extract all models and provider configs

feat(debug): add /debug/omniroute/db inspection endpoint

feat(models): expand /v1/models catalog with all live antigravity, nvidia and auto routes

Sep 19

fix(gateway): accept and pass body_override in handle_omniroute_proxy

feat(claude-code): add /messages endpoint routing and automatic claude model mapping to auto/smart

perf(logging): silence routine uvicorn access logs, router polling traces, and progress bars

fix(db): prevent WAL lock collisions, disable bundle regex patching, and add self-healing supervisor

feat(vault): dynamic token lookup and detailed status in vault_sync

feat(persistence): add /debug/persistence diagnostic and manual /debug/vault/backup endpoints

feat(persistence): bulletproof zero-loss persistence with cloud vault dataset sync and safe restart

Sep 18

feat(omniroute): dedicated space deployment, remove legacy projects and update credentials

Sep 12

Update gateway/utils.py

Update gateway/utils.py

Update gateway/jellyfin.py

Aug 31

fix(omniroute): import asyncio and json at top level to fix NameError on credential flush

fix(persistence): add continuous 10s database sync loop and immediate post-write atomic snapshot flush

fix(proxy): catch RemoteProtocolError and HTTPError with automatic retry in proxy_http_request

feat(oauth): add OAuth paste-credentials, keys capture, and direct SQLite vault sync

feat(vault): automatically capture and sync Web UI credentials to persistent vault and Hugging Face Space secrets

Aug 27

fix(omniroute): neutralize auth login rate limiting, extract client IPs, add static fallbacks

feat: completely remove Open WebUI and streamline multi-service container hub

Aug 15

fix(omniroute): intercept auto/ and omniroute/ model requests in gateway proxy to resolve directly to working connected provider models

fix(hermes): default HERMES_MODEL to antigravity/gemini-3.6-flash-medium and bypass failing public auto/ scrapers in gateway auto resolver

fix(hermes): point Telegram bot & Hermes agent API base URL to gateway port 8000 for auto model resolution; filter auto/ aliases

fix(omniroute): enable OMNIROUTE_ALLOW_UNAUTHENTICATED for internal gateway proxy and remove fake 200 unauthenticated response override on API routes

fix(omniroute): transparently return raw backend status codes and error bodies from OmniRoute port 20128

fix(hermes): implement multi-model trial loop over synced NVIDIA NIM and Antigravity provider models for 100% reliable chat responses

fix(hermes+omniroute): filter real NVIDIA NIM and Antigravity provider model IDs for automatic chat completion execution

fix(omniroute+hermes): wire OmniRoute API key sk-6646a5f2024f6318-d27ff7-f3e152c8 across all gateway services for authenticated model execution

fix(hermes): dynamically resolve auto model to synced NVIDIA NIM / Antigravity model IDs before posting to OmniRoute

fix(hermes): add automatic fallback to synced OmniRoute models if free pool returns empty choices

fix(hermes+omniroute): provide clear guidance in fallback response for connecting AI provider keys in OmniRoute dashboard

feat(hermes): connect Hermes Agent LLM brain to OmniRoute (20128/v1), add dynamic model routing, and enable external client API connection

Aug 14

fix(gateway+webui+hermes): intercept 401/403 status on omniroute routes to prevent openwebui auto-signout and validate choices array for telegram bot

fix(webui+hermes): pin WEBUI_SECRET_KEY to prevent session token invalidation and add usage object to chat completions for litellm compliance

fix(webui+hermes): unblock Open WebUI auths/signin route and update hermes .env and config.json api_base to port 8642 for telegram bot

fix(webui+hermes): remove manual BUILD_DIR override to fix Open WebUI 404 and set HERMES_API_BASE_URL to port 8642 to resolve telegram bot provider retries

fix(omniroute): add clean 200 OK handlers for favicon, provider models, and model test endpoints to eliminate console errors

fix(omniroute): add 401/403/502 chat completions fallback handler and intelligent POST router to ensure telegram and custom apps receive 200 OK

fix(stream): add proxy_buffering off, X-Accel-Buffering no headers, and sitecustomize monkeypatch to eliminate blank responses and sentence-transformers errors

fix(hermes): add rich response alias fields, sse streaming fallback, and async boot to prevent run failures in custom apps

style(logging): standardize neat structured iso-timestamped log formatting across all services

feat(hermes): add latency tracking, tool invocation counters, and structured telemetry logging

fix(routing): route /home and omniroute dashboard subpages to omniroute port 20128