7.4 billion tokens per month. 34 free LLM providers. 635 free model endpoints. All behind one /v1 endpoint, plus any custom OpenAI-compatible endpoint. Smart routing, automatic failover, encrypted keys. Personal experimentation only.
TypeScript
Commits on main
chore(catalog): V14 — disable Cerebras qwen-3-235b + llama3.1-8b ahead of 2026-05-27 deprecation (#79)
daf7fdc Tashfeen committed 6 months ago
feat(catalog): V13 — cross-provider refresh + HuggingFace router (May 2026) (#78)
f0ac1bd Tashfeen committed 6 months ago
chore(catalog): V12 — refresh OpenRouter :free pool (May 2026) (#77)
d5e919e Tashfeen committed 6 months ago
fix(router): treat 404 model-removed as retryable (#76, closes #66, credits @JammyJames1234)
8aea92c Tashfeen committed 6 months ago
feat(proxy): accept OpenAI multimodal array content envelope (#75, closes #69, credits @jhash)
f2ece37 Tashfeen committed 6 months ago
chore: drop orphan PDF + de-Raspberry-Pi-ify README/env (#74)
0b755d1 Tashfeen committed 6 months ago
fix(google): sanitize JSON Schema fields Gemini rejects (#65, credits @danscMax)
6929cd5 danscMax committed 6 months ago
fix(router): treat 413 Payload Too Large as retryable (#64, credits @danscMax)
839fe5a danscMax committed 6 months ago
feat(proxy): expose 'auto' as a virtual model in /v1/models (#62, closes #50, credits @mybropro)
2121550 Tashfeen committed 6 months ago
docs: DASHBOARD_ORIGINS + credit @nordbyte + Star History chart (#61)
8cb6550 Tashfeen committed 6 months ago
fix(security): require API key on loopback + restrict CORS to dashboard origins (#55)
8dd9130 Ricardo committed 6 months ago
fix(google): guard SSE JSON.parse so a bad chunk doesn't kill the stream (#47)
10bcfe2 Praveen Kumar Pranjal committed 6 months ago
feat(catalog): V11 — qwen3-235b bug fix, NVIDIA re-enabled (+9), Cerebras/Groq refresh (+4), +3 keyless aggregators (+7) (#46)
1d32108 Tashfeen committed 6 months ago
docs: credit @deadc for #42 and @zhangyu1324 for the Ollama Cloud request (#43)
c73c2a1 Tashfeen committed 6 months ago
feat(router): exhaust all keys before falling back + tests (#42, supersedes #33)
9b97219 Tashfeen committed 6 months ago
feat(catalog): integrate Ollama Cloud as a new platform (V10) (#41)
701e61c Tashfeen committed 6 months ago
security: phase 1 hardening (subset of #36, credits @jtbrennan-git) (#40)
8415820 Tashfeen committed 6 months ago
test(google): camelCase thoughtSignature fixture + credit @deadc for #32 (#34)
0441d25 Tashfeen committed 6 months ago
fix(google): preserve Gemini thought_signature in tool calls (#32)
130c395 Thiago Freitas committed 6 months ago
Fix server and Vite dev proxy port config from .env (#27)
b143fb3 Pham Quang Vinh VNU committed 6 months ago
fix: quality pack — crypto guard, session hash, idempotency test, integration tightening, comments (#26)
1e149e0 Tashfeen committed 6 months ago
chore: remove dead provider classes + fix orphan test fixtures + gitignore .bak (#25)
d1b3b4e Tashfeen committed 6 months ago