Platform Backend
General
This part of the Promethist documentation is for developers and technical users who connect other software to Promethist Studio, the application in which Promethist agents are built, tested, published and analyzed.
Using the Studio itself is described in the Platform User’s Guide. Integrating your own client application with the conversational runtime — the API and protocols agents talk through — is described in the Platform Core documentation.
Promethist for Claude
Promethist runs a hosted Model Context Protocol (MCP) server in every region. Claude connects to it and works in your Promethist workspace on your behalf: it can build and refine agents, manage their knowledge, interactive content, integrations and evaluations, run tests, read analytics and conversations, and manage the members of your accounts and projects — always with the permissions your Promethist account has.
- Promethist for Claude
-
What the connector and the Promethist plugin for Claude are, and what you need to use them.
- Connecting Claude
-
The connector address of each region, and how to connect Claude on the web, Claude Desktop, Cowork and Claude Code.
- Sign-In and Permissions
-
How Claude signs in to Promethist, how long the sign-in lasts, and what Claude may do with your roles.
- Tools
-
The tools the server offers, by area.
- Operating Guides
-
The guides Claude reads to learn how Promethist works.
- Files, Links and Conversation Data
-
Uploading files, downloading exports, and what conversation data Claude can read.
- Troubleshooting
-
What the messages you may see mean and what to do about them.
Regions
Promethist is operated in separate regions. An account, its projects, its agents and its data live in exactly one of them, and you sign in to the region your account is hosted in:
| Region | Studio | MCP server |
|---|---|---|
EU |
||
US |