claude-code-proxy vs claude-code-forge
Claude Code model access tools compared
claude-code-proxy translates Anthropic-format requests to OpenAI, Gemini, or Anthropic backends, while claude-code-forge configures Claude Code to use TensorBlock Forge. The proxy is a Python gateway with provider routing; Forge is a Shell-based setup utility that relies on Forge for model access.

claude-code-proxy: Connect Claude Code to Other Model Providers
A Python proxy that lets Anthropic-compatible clients such as Claude Code use OpenAI, Gemini, or Anthropic backends through LiteLLM. It suits developers who want to route Claude-style requests to another provider.

claude-code-forge: Connect Claude Code to Other AI Models
Shell scripts configure Claude Code to use TensorBlock Forge as its API provider, making Forge-supported models available through Claude’s interface. It suits users who already have Claude Code and a Forge account and want to switch among supported models.
| claude-code-proxy | claude-code-forge | |
|---|---|---|
| Language | Python | Shell |
| License | — | MIT |
| Stars | 3.8k | 28 |
| Forks | 494 | 8 |
| Last analyzed | Oct 9, 2026 | Oct 3, 2026 |
Key differences
- claude-code-proxy routes requests directly to OpenAI, Gemini, or Anthropic, while claude-code-forge routes them through TensorBlock Forge.
- claude-code-proxy translates requests and responses and supports streaming and non-streaming; claude-code-forge provides startup scripts that configure Claude Code.
- claude-code-proxy is written in Python and can run from source with uv or as a Docker container; claude-code-forge is written in Shell and provides scripts for Linux, macOS, and Windows.
- claude-code-proxy supports configurable model mappings and Google AI Studio or Vertex AI authentication; claude-code-forge uses editable default model identifiers and requires Forge credentials and provider keys.
- claude-code-proxy has 3.8k stars and 492 forks; claude-code-forge has 28 stars and 8 forks.
- claude-code-forge specifies an MIT license; claude-code-proxy metadata does not specify a license.
Choose claude-code-proxy if you…
- need a proxy for Anthropic-compatible clients that can route to OpenAI, Gemini, or Anthropic.
- want configurable Claude-style model mappings or streaming and non-streaming support.
- prefer a Python proxy that can run from source or in Docker.
Choose claude-code-forge if you…
- already use TensorBlock Forge and want to configure Claude Code to use it.
- want startup scripts for Linux, macOS, or Windows, with editable model identifiers.
- plan to use a self-hosted Forge deployment with Claude Code.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.