manifest.build

Manifest is an open source LLM gateway that gives AI agents and applications a single OpenAI compatible endpoint to reach many model providers. Users connect API keys, subscription plans such as Claude, ChatGPT or Gemini, local models through Ollama, LM Studio or llama.cpp, and custom providers, then define routing rules that send each request to the right model. Beyond routing, it provides automatic fallback to another provider when one fails, full request and response body logs, and cost tracking broken down by agent, key and provider, with budget alerts and notifications. Its Autofix feature repairs malformed requests on the fly before the provider returns an error, and it can run through the gateway or on its own. The site also announces an API Bot aimed at keeping an app working through upstream API breaking changes. The product is aimed at individual AI builders who want to get more from several subscriptions, teams that need to manage inference spend, and platforms that want fewer client errors. It can be self hosted with Docker under the MIT license or used as Manifest Cloud. The Free plan costs $0 per month and includes unlimited agents, all providers, Autofix, seven days of dashboard retention and community support, with 10,000 routed requests per month on Cloud and no limit when self hosted. Pro costs $19 per month, with unlimited routed requests, 365 days of dashboard retention and basic support. Enterprise is priced on request and adds on premise deployment, SSO, audit logs and an uptime SLA.