High-Performance LLM Proxy
Specifically built for Claude (Claude.ai) and Gemini (Google AI Studio, Google Vertex AI)
Core Advantages
Full-Featured Frontend
- Integrated React frontend providing a complete functional experience
Efficient Architecture
- Occupies one-tenth the resources compared to script language implementations, with ten times the performance, easily handling thousands of requests per second
- Event-driven design, decoupled logic, supports hot reloading and multiple configuration methods
- High-performance response caching supported by Moka technology
- Multi-threaded asynchronous processing based on Tokio and Axum
- Fingerprint-level Chrome simulation Rquest HTTP client
Intelligent Cookie Management
- Automatic classification and management of account status
- Fine-grained polling mechanism to maximize resource utilization
Full Platform Compatibility
- Rust static compilation, single binary deployment, no environment dependencies needed
- Native support for macOS/Android and other platforms
- Extremely low memory usage (only single-digit MB)
- No need for virtual machines or complex dependencies
Enhanced Features
- Built-in proxy server support (no TUN required)
- Concurrent cache request handling
- Gemini additional support:
- Google AI Studio and Google Vertex AI
- OpenAI compatible mode / Gemini format
- Painless HTTP Keep-Alive support
- Claude additional support:
- OpenAI compatible mode / Claude format
- Extend Thinking
- Stop sequences implemented on the proxy side
- Image attachment uploads
- Web search
- Claude Max
Claude Code 2.1.258 compatibility. Requests now identify as the current
client instead of the legacy 2.1.76 release. Billing attribution also joins
JavaScript UTF-16 code units 4, 7, and 20 before UTF-8 hashing, keeping the
cc_version suffix correct for emoji and surrogate-pair boundaries.
Portable GNU/Linux archives. Nix-built GNU binaries have their loader
path repaired before packaging, so the release ZIPs run outside the Nix
store on supported distributions instead of failing with “No such file or
directory”.
Container publication. Multi-architecture image publishing now preserves
the existing sha-<7> tags, produces real amd64/arm64 manifest lists, and
avoids publishing images from manual workflow dispatches.
Changes
Reproducible release builds. Linux GNU, musl, Android, frontend and
distroless container outputs now use pinned Nix definitions. nix develop
provides the matching stable/wasm/nightly/trunk toolchain, while Windows and
macOS retain their native build paths. The HTTP stack and other dependencies
were refreshed, including the stable wreq 0.16 release.
Faster, verifiable CI. Linux targets and both container architectures now
build in parallel; exact feature sets, shared frontend store paths and
minimal per-target toolchains avoid redundant compilation. Workflow YAML and
the image-push/tag scheme are also checked locally through xtask before
publication.
