THE STANDARD
What the Model Context Protocol is.
MCP is the open standard that lets an AI assistant use an outside tool. It is how your Twin is read.
Last verified
What is the Model Context Protocol?
An open standard that lets an AI assistant use an outside tool. Instead of each assistant inventing its own plugin system, one protocol describes how a tool announces what it can do and how the assistant calls it.
Who controls it?
No single company. It was donated to the Agentic AI Foundation under the Linux Foundation in December 2025.
Which assistants support it?
Support varies by assistant and by plan. The dated position for each is listed below.
Where it came from
The Model Context Protocol was released as an open standard around 25 November 2024 under an MIT licence.
Verified . Primary source.
MCP was donated to the Agentic AI Foundation under the Linux Foundation on 9 December 2025, co-founded with Block and OpenAI, with platinum members including AWS, Google, Microsoft, Cloudflare and Bloomberg. It is no longer any single company's protocol.
Verified . Primary source.
Anthropic's 9 December 2025 ecosystem update cited more than 10,000 active public MCP servers and more than 97 million monthly SDK downloads across Python and TypeScript.
Verified . Primary source.
Who supports it, by plan
Free with one connector, Pro, Max, Team and Enterprise.
Verified . Primary source.
Plus and Pro for read and fetch through Developer Mode. Business, Enterprise and Edu for write-capable connectors. Not available on free accounts.
Verified . Primary source.
Pro, Max and Enterprise.
Perplexity's own documentation has been inconsistent about availability, so check in your account.
Verified . Primary source.
All plans including free, launched 2 September 2025.
Verified . Primary source.
Not verified. We have not been able to confirm this from xAI's own documentation.
Verified . Primary source.
No clear route for a user to add a custom connector.
Verified . Primary source.
No clear route for a user to add a custom connector.
Verified . Primary source.
Why it matters for memory
A memory layer is only useful if the assistant you happen to be using can read it. MCP is what makes that possible without a separate integration for every assistant. How to connect your Twin.