Condens Remote MCP Server
Remote MCP server for connecting to your Condens UX repository.
Que pouvez-vous faire avec Condens Remote MCP Server ?
- Published Research Access - Search and retrieve published Condens Artifacts, including insights, reports, usability issues, and other research outputs
- Permission-Aware - Users only see data they already have access to in Condens
- Client Access Control - Admins control which MCP clients are available and can scope access by role, user group, or individual user
- OAuth 2.0 Authentication - Uses OAuth 2.0 authorization with support for dynamic client registration and PKCE
- Streamable HTTP Transport - Exposes the MCP server over Streamable HTTP at
/mcp
Installation
Ajoutez ceci à la configuration MCP de votre client (Claude Desktop : claude_desktop_config.json ; Cursor : .cursor/mcp.json ; VS Code et les autres utilisent la même forme « mcpServers »).
claude mcp add --transport http condens-remote-mcp-server 'https://{baseUrl}/mcp'
{
"mcpServers": {
"condens-remote-mcp-server": {
"type": "http",
"url": "https://{baseUrl}/mcp"
}
}
}
Documentation
Tirée du README du projet sur GitHub, actualisée chaque semaine. Les liens ouvrent GitHub.
Sommaire
Condens MCP Server
Overview
The Condens MCP Server lets you connect MCP-compatible AI assistants, IDEs, and agent platforms to your Condens research repository. Once connected, you can search across your validated customer insights in natural language directly from the LLMs you already use, or let agents and workflows pull from research findings so whatever they produce is grounded in real customer knowledge.
Features
- Published Research Access - Search and retrieve published Condens Artifacts, including insights, reports, usability issues, and other research outputs
- Permission-Aware - Users only see data they already have access to in Condens
- Client Access Control - Admins control which MCP clients are available and can scope access by role, user group, or individual user
- OAuth 2.0 Authentication - Uses OAuth 2.0 authorization with support for dynamic client registration and PKCE
- Streamable HTTP Transport - Exposes the MCP server over Streamable HTTP at
/mcp
Available Tools
The server currently provides these read-only MCP tools:
- List Workspaces - Lists Condens Workspaces accessible to the authorized user
- List Published Artifacts - Lists accessible published Artifacts, optionally filtered by Workspace ID or Artifact IDs
- Get Artifacts - Retrieves Markdown content of Artifacts by ID
- Search Condens - Searches published Artifacts for content relevant to a user query and returns matching Artifact IDs
Getting Started
1. Set up the MCP client in Condens
Admins create and configure the MCP client in Condens under Settings > Account > API & MCP Server, including which users, roles, or groups are allowed to use it.
2. Add Condens MCP to your tool
In your MCP-compatible tool, add the Condens MCP server URL. Your URL depends on your Condens server region:
https://api.condens.io/mcpfor EUhttps://api-us.condens.io/mcpfor US
If you're unsure about your Condens server region, Admins can find the designated MCP URL in the API & MCP Server page.
Condens includes preconfigured MCP clients for Claude and ChatGPT. See the client-specific guides below for step-by-step instructions.
3. Authorize the connection on user-level
Each user authorizes the connection for their own Condens account via OAuth. Authorized non-admin users can find setup instructions in: Settings > Integrations > MCP Client > Show setup instructions.
Authentication uses OAuth 2.0 with Dynamic Client Registration. Compatible MCP clients register themselves during the authorization flow, so Condens does not provide manually generated client IDs or client secrets to copy into your MCP client.
Condens MCP currently supports remote MCP over Streamable HTTP. It does not support local stdio servers or SSE.
For full setup instructions, view our step-by-step setup guide.
Client-specific guides:
Availability
Condens MCP is available during the free 15-day trial and on Business and Enterprise plans.
Business plans can connect Condens to preconfigured AI assistants such as Claude and ChatGPT.
Enterprise plans can additionally connect Condens to custom MCP clients, including custom or company LLMs, automation agents, design assistants, knowledge assistants, and coding assistants.
Some MCP clients may require their own paid plan, such as Claude Pro or higher, or ChatGPT Plus or higher.
Security & Access Control
Condens Admins control which MCP clients are available for their organization. For each MCP client, Admins can configure who is allowed to use it by selecting roles, user groups, or individual users.
Only authorized users see the MCP client in their Integrations settings and can complete the OAuth authorization flow. Each user authorizes the MCP client for their own Condens account, so access remains scoped to that user's existing Condens permissions.
Admins can remove or deactivate an MCP client at any time. They can also change which users, roles, or groups have access; those changes limit who can continue using the MCP connection.
Support
For questions about Condens MCP setup, contact Condens via hello@condens.io
Questions fréquentes
Qu'est-ce que Condens Remote MCP Server ?
Remote MCP server for connecting to your Condens UX repository.
Comment installer Condens Remote MCP Server ?
Il fonctionne comme serveur distant. Dans Claude Code : claude mcp add --transport http condens-remote-mcp-server 'https://{baseUrl}/mcp'. Pour les autres clients, collez la configuration JSON de cette page dans leurs réglages MCP.
Faut-il une clé API pour Condens Remote MCP Server ?
Son entrée au registre n'indique aucune clé ni secret obligatoire. Certaines fonctions peuvent tout de même demander un compte auprès du service auquel il se connecte.
Condens Remote MCP Server est-il open source ?
Oui. Son code est sur GitHub sous licence MIT.
Qui maintient Condens Remote MCP Server ?
Il est publié par condens. AI.tn le référence sans le maintenir ni le recommander.
Issu du registre MCP officiel. Être listé n'est pas une recommandation : vérifiez le code et les permissions d'un serveur avant de lui donner accès à vos données.