Manifest is unclaimed —

Manifest

The Open Source LLM Router for AI Agents·manifest.build

From $19/mo · Free plan · Open source See pricing ↓

Manifest is an open source LLM router that connects AI agents and harnesses to any LLM provider in seconds. It routes requests to subscription-based providers, pay-per-token APIs, local models, or custom endpoints through a single OpenAI-compatible or Anthropic-compatible endpoint. It includes model fallbacks, an Auto-fix feature that repairs broken provider requests on the fly, and observability into request, token, and cost data. It can be used as a hosted Cloud service or self-hosted via Docker.

What it's for

Connecting AI agents and coding assistants to multiple LLM providers through one endpointUsing existing AI subscriptions (Claude Pro, ChatGPT Plus, Gemini Advanced) with agentsAutomatically retrying/fallback when a provider is down or rate-limitedRepairing broken or invalid provider requests before they reach the agent (Auto-fix)Routing requests to local models (Ollama, LM Studio, llama.cpp) instead of the cloudTracking spend and usage per agent, key, and providerSetting budgets, rate limits, and custom routing tiers per agent

Features 15

Auto-fix: repairs broken or invalid provider requests on the fly
Bring-your-own API keys and OAuth tokens for subscription providers
Budget alerts and notifications
Cost visualization by agent, key, and provider
Custom provider support (OpenAI-compatible endpoints)
Error catalog classifying common provider errors
Full request/response body logging
Local model routing (Ollama, LM Studio, llama.cpp)
Model discovery with per-model cost data (GET /v1/models?cost=true)
Model fallbacks across providers when one fails
OpenAI-compatible and Anthropic-compatible API endpoints
Routing rules and custom tiers based on request headers
Self-hosting via Docker
Single endpoint (LLM Gateway) to access all providers and models
Support for 500+ models across providers

At a glance

free tieropen sourceself-hostableAPI

TypeWeb app

DeploymentHybrid

PlatformsWeb, Windows, macOS, Linux

ForAI builders, teams & startups, platforms & providers, developers

CompanyManifest

Integrations

DockerGoogle GeminiHugging Face Inference ProvidersCraft AgentAnthropic SDKMoonshot KimiClaude CodeNanobotMistralGroqOpenAIDiscordNVIDIA NIMDeepSeekllama.cppGitHubHermes AgentOllamaLM StudioAnthropicChatGPTMiniMaxClaudeOpenAI SDKOpenCodeAlibaba Cloud QwenVercel AI SDKOpenClawLangChain

Complianceself-reported

ISO 27001SOC 2 Type IIHIPAA

Resources

Pricing

Enterprise plans (Cloud and Self-hosted) are quote-based ('Talk to sales'). Free Cloud plan pauses requests beyond 10,000/month until the next billing cycle. A promotional offer of $25 Gemini credit is given for a 15-minute discovery call.

Free (Manifest Cloud)Free

For prototypes and small projects. Requests beyond 10,000/month are paused until next billing cycle.

10,000 routed requests / month; 7-day dashboard retention

Pro (Manifest Cloud)$19/mo

For production projects.

Unlimited routed requests; 365 days dashboard retention; not suited for teams

Enterprise (Manifest Cloud)Quote

Quote-based, 'Talk to sales'.

For scaling projects and large-scale teams; quote-based

Open Source (Self-hosted)Free

Run on your own infrastructure; free with no restrictions.

No usage limitations stated

Enterprise (Self-hosted)Quote