Manifest — Smart LLM Router That Cuts Costs 70%

Introduction

Manifest is a smart LLM router with 4,200+ GitHub stars. It sits between your app and LLM providers, scores each request across 23 dimensions (<2ms), and automatically routes to the cheapest capable model — saving up to 70% on cost. Supports 300+ models across 13+ providers, with automatic failover and budget controls. Ideal for teams running production LLM apps that want to optimize API spend automatically.

Manifest — Smart LLM Cost Optimization

How It Works

1. Request arrives → 23-dimension complexity scoring (<2ms)
2. Pick the cheapest sufficiently-capable model
3. Route to the chosen provider
4. On failure, automatically switch to a fallback model

Core Features

• 300+ models across 13+ providers
• 23-dimension scoring, <2ms latency
• Up to 70% cost savings
• Automatic failover
• Budget controls
• Transparent decision dashboard

FAQ

Q: What is Manifest? A: A smart LLM router that scores requests across 23 dimensions and routes to the cheapest capable model — saving up to 70% on API cost.

Q: Is it free? A: The core router is open source (MIT) and self-hostable.