Our take
Quietno commits in the last 90 days and no release in a year- Activity
- 19
- Community
- 58
- Issues
- 21
- Pull requests
- 20
+7 stars in the last 30 days
This is a focused way to bridge Anthropic-compatible clients to multiple providers, but its quiet recent activity makes it worth validating before relying on it in production.
Good fit if
- You need Claude Code or another Anthropic-compatible client to use OpenAI, Gemini, or Anthropic backends.
- You can run a Python service or Docker container and configure provider credentials yourself.
- You are comfortable assessing a project with no recent commits or releases, but contributions from multiple people rather than one dominant maintainer.
Look elsewhere if
- You need a project with recent release activity or demonstrated recent issue and pull-request handling.
- You need a documented test suite as part of your adoption requirements.
- You need a clearly stated license before using or distributing the software.
All health signals
| Last commit | 2026-06-23 (4 months ago) |
|---|---|
| Commits, last 90 days | 0 |
| Releases, last 12 months | 0 |
| Contributors | 12 (top contributor: 39% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 19 months |
Checked on 2026-10-09 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
3,764Forks
485Change since Nov 16, 2026: +2 stars·0 forks
Overview
claude-code-proxy accepts requests in Anthropic's API format, translates them through LiteLLM, and returns Anthropic-formatted responses. This lets clients such as Claude Code use OpenAI or Gemini models, or pass requests through to Anthropic.
It is aimed at developers who want to keep an Anthropic-compatible client while choosing or switching the model provider behind it. Model aliases such as Haiku and Sonnet can be mapped to configured models, with provider selection controlled by environment variables.
Key Features
- Exposes an Anthropic-compatible API endpoint for clients such as Claude Code.
- Routes requests to OpenAI, Gemini, or Anthropic through LiteLLM.
- Maps Haiku and Sonnet aliases to configured models for OpenAI and Google providers.
- Supports Google AI Studio keys or Vertex AI authentication with Application Default Credentials.
- Handles streaming and non-streaming responses.
- Can run from source with
uvor as a Docker container.
Use Cases
- Claude Code users who want to route requests to OpenAI or Gemini models.
- Developers who want to switch provider or model through configuration rather than changing their Anthropic-compatible client.
- Teams using Vertex AI that want to connect a Claude-compatible client through a locally hosted proxy.
- Developers who want proxy middleware while continuing to use Anthropic models directly.
What you need
Detected in the repository
- Python >=3.10 (from pyproject.toml)
- A Dockerfile, so it can run in a container
- Automated checks on GitHub Actions
Getting Started
After cloning the repository and configuring .env, run:
uv run uvicorn server:app --host 0.0.0.0 --port 8082
Then point Claude Code at the proxy with ANTHROPIC_BASE_URL=http://localhost:8082 claude. See the README for provider configuration and Docker instructions.
License in Plain Words
No standard license was detected. Without a license, the default copyright rules apply: you may read the code, but not reuse it. Check the repository or ask the authors.
Alternatives
- claude-code-forge: It configures Claude Code to use TensorBlock Forge specifically, while claude-code-proxy routes Anthropic-compatible requests to multiple backends through LiteLLM.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| claude-code-proxy | Python | – | 3.8k | Quiet |
| claude-code-forge | Shell | MIT | 28 | Inactive |
Considerations
The project is quiet by the supplied health data: it has had no commits in the last 90 days and no release in over a year. Multiple contributors have contributed, and work is not concentrated in one person, but the supplied metrics show no recent issue closures or merged pull requests. The repository data does not specify a license, and does not indicate a test suite. Google Vertex AI use also requires a configured cloud project and region; other providers require their corresponding credentials. Check model mappings against your intended provider and model, since Google routing depends on the proxy's recognized Gemini models.
Found this useful?
Share it with someone who would like claude-code-proxy.