Manus Flex: Bring Your Own Model API Key

Manus Flex (beta) lets you connect an API key from a supported provider so the language model runs on your account while Manus continues to handle planning, tools and execution environments. It is for users who already have an account, credits or a contract with an inference provider.

Supported providers#

As of the October 2026 docs: OpenAI, Anthropic, Google AI Studio, xAI, Meta, OpenRouter, Fireworks and Modal.

Billing split#

Item	Who bills it
Model inference (the model that runs the agent)	Your inference provider, through your API key
Cloud sandbox (temporary)	Covered by Manus at no additional cost
Images, video, Cloud Computer, web app hosting and other paid features	Manus credits

Flex is useful if you already have favorable rates, preferred models or compliance requirements around model choice. Manus notes that third-party model performance may vary and is not guaranteed. See Manus pricing for credit costs and Manus security before connecting keys.

Sources & Further Reading#

Claim / Topic	Source	Date
Flex beta, supported providers, billing split	Manus docs: Manus Flex	2026-10-30

Frequently asked questions

What is Manus Flex?

Manus Flex, in beta, lets you connect an API key from a supported provider so the language model runs on your account while Manus still handles planning, tools and execution. Supported providers are OpenAI, Anthropic, Google AI Studio, xAI, Meta, OpenRouter, Fireworks and Modal. Images, video and Cloud Computer still use Manus credits.

Which providers work with Manus Flex?

OpenAI, Anthropic, Google AI Studio, xAI, Meta, OpenRouter, Fireworks and Modal, according to the Manus Flex docs.

What does Manus still charge for with Flex?

Model inference is billed by your provider and temporary sandboxes are covered by Manus. Images, videos, Cloud Computer, web app hosting and other paid features still use Manus credits.