awesome-repositories.com
Blog
MCP
awesome-repositories.com

Discover the best open-source repositories with AI-powered search.

ExploreCurated searchesOpen-source alternativesSelf-hosted softwareBlogSitemap
ProjectMCP serverAboutHow we rankPress
LegalPrivacyTerms
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
yetone avatar

yetone/magpie

0
View on GitHub↗

Magpie

Magpie is a local gateway for coding agents. It points tools such as Claude Code, Codex and Gemini CLI at models from any provider by writing each tool's own config file, so a user picks which model each agent runs on without editing configs by hand. The gateway translates between OpenAI Chat, OpenAI Responses, Anthropic Messages and Gemini formats while preserving streaming, tool calls and reasoning, so any client that accepts a base URL can reach any model.

Several features set it apart. Existing subscriptions, such as Claude, ChatGPT or Copilot plans, can be exposed as providers for every other agent, with multiple accounts per plan. Models, accounts and keys can be grouped under one name and balanced by order, rotation, usage, pace or renewal timing, with automatic failover when a quota runs out and optional routing by intent. A local masking layer replaces keys, tokens and personal data in prompts with placeholders before they leave the machine, restores the real values in replies, and lets users audit what was altered. Magpie also tracks tokens, cost, quota windows and sessions per key and account, and can export telemetry to OpenTelemetry or Langfuse.

Beyond routing, Magpie writes instructions, MCP servers, skills and subagents once into each agent's own file format. It also supports saved setup profiles and encrypted settings sync over WebDAV, S3 or a private repository. JavaScript plugins and middleware can add providers, search backends and new agents, or rewrite, reject and inspect requests. Other features include image and video generation tools, web search for models that lack it, embeddings and rerank endpoints, and sharing the gateway across a network with per-client keys and spending limits.

Magpie can be run from the app, a menu bar icon, a terminal screen or a browser console. A Docker image is also available that keeps all state in a single volume and protects access with a gateway key.

AI search

Explore more awesome repositories

Describe what you need in plain English — the AI ranks thousands of curated open-source projects by relevance.

Start searching with AI
usemagpie.ai
↗

Features

  • LLM Gateways - Runs a local gateway that routes coding agents to any model provider and translates between OpenAI, Anthropic and Gemini formats, preserving streaming, tool calls and reasoning.
  • Model Routing for Agents - Points coding-agent tools such as Claude Code, Codex and Gemini CLI at models from any vendor through one local gateway without hand-editing configs.
  • Agentic Traffic Routing - Points coding agents such as Claude Code, Codex and Gemini CLI at a local gateway so any chosen provider model serves them.
  • Agent Model Switchers - Opens a searchable list of every provider model and rewrites the chosen agent's config in place, keeping comments and formatting.
  • Model Group Request Routing - Groups models, accounts and keys under one name and spreads requests by order, rotation, usage, pace or soonest renewal.
  • Multi-Provider LLM Gateways - Runs a local gateway that lets any base-URL tool reach models from any provider through format conversion with streaming and tool calls.
  • Prompt Privacy Filters - Swaps keys, tokens and personal data for placeholders locally before prompts leave the machine, then restores the real values in replies.
  • Quota Failover Routing - Switches to the next account or model automatically when a quota runs out, hiding the error from the agent.
  • Subscription Reuse Across Agents - Exposes signed-in Claude, ChatGPT or Copilot plans as model providers for every other agent, with several accounts per plan.
  • Format Translators - Converts between OpenAI Chat, OpenAI Responses, Anthropic Messages and Gemini requests, covering streaming, tool calls and reasoning.
  • Agent Subscription Discovery - Reads existing sign-ins from agent tools, refreshes tokens, and shares each subscription's models with every other agent.
  • Instruction and Tool Syncing - Writes instructions, MCP servers, skills and subagents once into each coding agent's own file format and removes only what it added.
  • Model Provider Integrations - Offers presets for many vendors and any compatible endpoint, pulls model lists and reasoning tiers, and imports settings from other tools with per-key proxies.
  • OpenAI-Compatible Endpoints - Serves OpenAI, Anthropic, Gemini and Ollama-compatible endpoints so any client can connect by URL, with optional gateway-key authentication.
  • Provider Account Rotation - Rotates requests among multiple accounts of a subscription provider and starts usage windows early on a schedule.
  • Web Search Tools - Searches on behalf of models that lack web search using ordered search services with multiple keys, and exposes search and page fetch tools to any agent.
  • Sensitive Data Redaction - Replaces secrets and personal data such as keys, tokens, emails and card numbers with placeholders before requests leave the machine, restoring them in replies.
  • Gateway Plugins and Middleware - Supports JavaScript plugins that add logins, providers and models, plus middleware that can rewrite, reject or inspect requests, events and responses.
  • Account-Based Load Balancers - Orders multiple subscription accounts or keys per request by remaining quota and rate-limit state, keeping a conversation on its account to preserve cache.
  • LLM Provider Failovers - Groups models, accounts and keys under one name, balances requests by order, rotation or usage, and fails over when a quota runs out.
  • Shared Local Network Access - Opens the local gateway to other machines using named client keys with daily, weekly or monthly token and cost limits.
  • Secret Masking - Replaces keys, tokens and personal data in prompts with placeholders locally, restores real values in replies, and lets users audit what changed.
  • Gateway Middleware - Runs JavaScript plugins that add providers or act as middleware, rewriting, rejecting or inspecting requests and responses through the gateway.
  • Provider Plugins - Installs provider plugins from published packages that sign in to subscriptions, list models and make requests on a downloaded JavaScript runtime.
  • AI Usage Quota Tracking - Shows account quota windows, tokens, spend in chosen currencies and per-session cost, with billing breakdowns and telemetry export.
  • Subscription Plan Quota Checks - Pools several accounts and plans behind named groups that rotate, balance load and fail over automatically when a quota runs out.
  • Cost and Token Trackers - Reports tokens, cache use, context fill, per-key quotas and estimated cost by account, with session history and observability export.
31,791 stars·2,793 forks·Go·MIT·11 views

Star history

Star history chart for yetone/magpieStar history chart for yetone/magpie

How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.

Frequently asked questions

What does yetone/magpie do?

Magpie is a local gateway for coding agents. It points tools such as Claude Code, Codex and Gemini CLI at models from any provider by writing each tool's own config file, so a user picks which model each agent runs on without editing configs by hand. The gateway translates between OpenAI Chat, OpenAI Responses, Anthropic Messages and Gemini formats while preserving streaming, tool calls and reasoning, so any client that accepts a base URL can reach any model.

What are the main features of yetone/magpie?

The main features of yetone/magpie are: LLM Gateways, Model Routing for Agents, Agentic Traffic Routing, Agent Model Switchers, Model Group Request Routing, Multi-Provider LLM Gateways, Prompt Privacy Filters, Quota Failover Routing.

Which projects share features with yetone/magpie?

Projects with overlapping indexed features include: ogx-ai/ogx — OGX is a self-hostable API server that exposes OpenAI-compatible endpoints for chat completions, responses, and… tashfeenahmed/freellmapi — freellmapi is a self-hosted gateway that combines the free tiers of many language model providers behind one local,… nvidia-nemo/switchyard — Switchyard is a Rust library for routing large language model requests. For each request it chooses which model and… router-for-me/easycliproxyapi — EasyCLIProxyAPI is a desktop application for installing, configuring, and controlling a local LLM API proxy. It puts… getagentseal/codeburn — CodeBurn is a local-first tracker, analyzer, and optimizer for AI coding spend. It reads the session files that coding… linuxhsj/openclaw-zero-token — OpenClaw Zero Token is a self-hosted gateway and assistant platform that lets people chat with browser-based AI…

Projects sharing features with Magpie

These projects share indexed features with Magpie. Shared tags can include platform or build tooling; verify the primary use case before treating a result as a replacement.
  • ogx-ai/ogxogx-ai avatar

    ogx-ai/ogx

    8,454View on GitHub↗

    OGX is a self-hostable API server that exposes OpenAI-compatible endpoints for chat completions, responses, and embeddings, and routes them to many local or hosted model backends. It also accepts requests from Anthropic Messages and Google GenAI client libraries, translating them into one native format so teams can keep the client libraries they already use. The same core can run as an HTTP server reachable from any language or as a library imported into a Python program. What sets it apart is that agent orchestration runs on the server. Tool calling, connections to MCP tool servers, file sea

    Python
    View on GitHub↗8,454
  • tashfeenahmed/freellmapitashfeenahmed avatar

    tashfeenahmed/freellmapi

    48,827View on GitHub↗

    freellmapi is a self-hosted gateway that combines the free tiers of many language model providers behind one local, OpenAI-compatible endpoint. Applications talk to a single address with a single bearer token, while the router chooses a provider and model for each request and retries on another when one errors or hits its limit. Provider keys are stored encrypted in a local database, and per-key rate limits are tracked against per-minute and per-day ceilings. Beyond the OpenAI-style format for chat, completion, embedding, image, video, and speech requests, the router also accepts Anthropic, G

    TypeScript
    View on GitHub↗48,827
  • nvidia-nemo/switchyardNVIDIA-NeMo avatar

    NVIDIA-NeMo/Switchyard

    3,494View on GitHub↗

    Switchyard is a Rust library for routing large language model requests. For each request it chooses which model and provider should respond, balancing answer quality, cost, and response time rather than relying on one fixed model. It is aimed at coding assistants and agent harnesses, where it picks the best available model for each step and can show the chosen model in a tool's status line. Routing decisions come from composable strategies. A language model can judge whether a cheaper, more efficient model can handle a task, and that verdict selects the model tier. Execution signals, such as

    Rust
    View on GitHub↗3,494
  • router-for-me/easycliproxyapirouter-for-me avatar

    router-for-me/EasyCLIProxyAPI

    3,494View on GitHub↗

    EasyCLIProxyAPI is a desktop application for installing, configuring, and controlling a local LLM API proxy. It puts the proxy core's status, process ID, listening port, and settings in one window, and lets you start, stop, restart, and refresh the core. It installs and updates both the core and the app itself, including offline installs and selectable mirror sources with automatic fallback. The proxy acts as a local gateway that combines several upstream provider accounts and endpoints behind one address, converting requests and responses between OpenAI, Claude, and Gemini API formats. You c

    TypeScript
    View on GitHub↗3,494
Compare all 30 related projects→