Connect tool servers
Attach external MCP servers without turning Lurus into a REST API for custom applications.
Model Context Protocol
External tool servers. Directly in Lurus Chat.
Connect MCP servers to Lurus and make external tools available to your AI agents in a controlled way – from tool discovery to confirmation dialogs.
Overview
MCP brings your tools into the AI workflow.
Lurus connects external tool servers via Streamable HTTP and exposes discovered tools safely in chat.
Discover tools automatically
Lurus reads available tools and makes them available to chat when they match the task.
Execute with control
Read-only tools can run directly, while write actions require clear confirmation.
How it works
From connection to execution.
MCP stays technically flexible while remaining controllable for teams.
Connect a remote MCP server
Add the server URL and optional headers. Lurus uses Streamable HTTP and encrypts sensitive headers.
Discover tools
Tool discovery makes available actions visible and keeps them current for chat.
Use them in chat
The AI identifies suitable tools in conversation and executes them according to read-only status and team policy.
Security & control
Built for business workflows.
MCP extends Lurus without giving up control over data flows and actions.
Streamable HTTP
Remote MCP servers are connected through the supported Streamable HTTP transport.
Encrypted headers
Authentication headers are stored encrypted and are not exposed visibly in the UI.
Team policies
Admins can allow or restrict integrations across the team.
Confirmation for write access
Non-read-only tools trigger explicit confirmation before actions are executed.
Use cases
Where MCP is especially strong.
Internal knowledge systems
Connect documentation, knowledge bases or internal search systems as tool servers.
Developer and documentation tools
Expose technical tools such as context servers, documentation search or issue systems.
CRM and project management
Access business data and control actions with confirmation for write operations.
Data sources and automations
Connect custom automations without building every action as a native integration.
FAQ
Questions about MCP
Ready for controlled tool servers?
Start with Lurus and bring external tools safely into your AI workflows.
MCP-ready Team policies Confirmations for write access