Manifest LLM Gateway

What Manifest LLM Gateway does

Manifest LLM Gateway connects your own provider API keys, model subscriptions, local models and custom OpenAI- or Anthropic-compatible endpoints to a single OpenAI-compatible endpoint. Sending "model": "auto" lets Manifest pick the model for each query. The README states support for 300+ models through 35 built-in provider connections and 18 subscription flows, with catalogs discovered when credentials are connected.

Beyond routing, it keeps full-body logs for successful and failed requests, tracks spend with notifications and limits, and falls back to other models when a query fails. It runs as a hosted service at app.manifest.build or as a self-hosted Docker image, with deployment guides for Railway, Render, DigitalOcean, AWS, GCP, Fly.io, Coolify, Easypanel, Heroku, Koyeb and Apple Containers. The older npm-based self-hosting path is no longer supported.

Key features

• One OpenAI-compatible endpoint for API keys, subscriptions and local models
• Custom routing across providers, with an auto model option
• Full-body logs for successful and failed requests
• Cost tracking with notifications and limits
• Fallback to other models when a query fails
• Hosted cloud version or self-hosted Docker image

When to use it

• Point an agent at one endpoint while switching between provider keys and subscriptions behind it
• Track what each agent request costs and set spending limits
• Fail over to a second model when a provider returns errors
• Self-host the gateway on a VPS or PaaS using one of the deployment guides

Who it is for: Developers running AI agents who want one routing, logging and cost-control layer in front of several model providers.

How it fits with Hermes Agent

Tagged hermes-agent and openclaw on GitHub; the README describes it as a gateway for AI agents and apps in general rather than a Hermes-only tool.

FAQ