Manus Flex

Lets you run Manus on your own API key from a supported model provider, for Manus users who already pay for AI model usage elsewhere.

You connect an API key from a supported inference provider, pick a primary model and a reasoning-effort setting, and Manus uses that model for its agent work
Manus still supplies the agent system around the model: planning, tools, execution environments, web app hosting, and databases
Model usage is billed directly by your provider. Other services and infrastructure used in a Manus task still use Manus credits
Launch providers are OpenRouter, Fireworks, and Modal, through a new Flex Inference Partner Program that is meant to add more
Works inside the projects, automations, and agent workflows you already run in Manus