Back to directory

aibridge

An HTTP gateway that sits between AI clients and the Anthropic and OpenAI APIs, recording token usage, prompts and tool calls per user and optionally injecting a centrally managed set of MCP tools.

aibridge

aibridge proxies requests to the Anthropic messages endpoint and the OpenAI chat completions endpoint, forwarding an authenticated actor's traffic to the upstream provider while an interceptor records token usage, prompts and tool invocations through a pluggable Recorder interface. Passthrough routes such as listing models and counting tokens are reverse-proxied directly without interception.

When MCP support is enabled, the gateway connects to configured MCP servers, centrally defines their tools with allowlist and denylist regex filtering, injects them into requests with a shared prefix, and invokes them in an inner agentic loop when the model selects one. It exposes Prometheus metrics for interception counts, in-flight requests, duration, token usage, prompt counts, injected tool invocations and passthrough counts. The project is now in maintenance mode, with further development continuing in the coder/coder repository.

More in LLM gateways

See all 106
  • aihappey-ai

    github.com

    A stateless, multi-provider .NET AI gateway exposing normalized OpenAI- and Anthropic-compatible endpoints, managed agents and skills across a very large catalog of models.

  • Unified AI System

    happy520ai.github.io

    A self-hosted, protocol-first AI gateway and agent control plane for OpenAI, Anthropic, Gemini, MCP and A2A, turning a rough request into a structured, reviewable prompt before it runs.

  • Subroute

    github.com

    A local AI gateway and desktop control desk that gives coding tools one steady endpoint, so switching between AI subscriptions, API providers and local models doesn't mean reconfiguring every client.