Spaces:
Sleeping
Sleeping
Commit History
-
Configure for Hugging Face Space (Docker): port 7860, non-root user, secret-driven config via run.py
fd2eea3
-
fix: surface BardErrorInfo as error instead of silent null content
When Gemini upstream returns BardErrorInfo (e.g. [1060] from Docker bridge networks), the server now returns a proper 502 error with the error code instead of silently returning content: null. Also adds Docker host networking note to README. Closes #249fb5cad
-
docs: fix outdated claims in README
- httpx is required for streaming, not stdlib-only - Pro routing needs Gemini Advanced (paid), not free account - Add pip install httpx to Quick Startb866ca0
-
Support authenticated Gemini account routing
ethansunqing04b9d27
-
unify acknowledgements and friend links format
c1c2d1f
-
Add GenericAgent credit
ee9ada5 -
docs: add limitations section, clarify no image support
- Add "Limitations" section to both README.md and README_CN.md - Clearly state: no multimodal, no real Pro, single-turn, rate limits - Created issue #12 with full technical explanation02697fb
-
Merge remote-tracking branch 'origin/main'
HouYunFei8ae5a8e -
feat(server): 添加 API 密钥认证和 Docker 部署支持
HouYunFei +1: CodexCo-Authored-By: Codex <noreply@openai.com>f39fd46 -
feat: add Google native API support for Gemini CLI compatibility
LiangLliuAdd /v1beta/models endpoints (generateContent, streamGenerateContent) so that Gemini CLI can connect directly via GOOGLE_GEMINI_BASE_URL.f805ec8 -
feat: add proxy support (--proxy / config / env var)
Users behind firewall can now configure HTTP proxy via: - CLI: --proxy http://127.0.0.1:7890 - config.json: {"proxy": "http://..."} - env: HTTPS_PROXY=http://...cda0962 -
docs: add logo to readme
cd31d7a -
docs: add cookie guide, cross-link CN/EN readme
88cedd4 -
init: gemini-web2api v1.0.0
OpenAI-compatible API server for Google Gemini web interface. Zero auth, cross-platform, single file, no dependencies.504aef8