DearTech-OS for Model Context Protocol (MCP)
DearTech-OS exposes the full company knowledge graph as a Model Context Protocol (MCP) server at mcp.deartech.co. Any AI client that speaks MCP can connect: Claude, ChatGPT, Cursor, Codex, custom agents built on the MCP SDK, and the growing list of tools that adopt the protocol. Build the integration once on your end, get every MCP-compatible tool for free.
Last updated:
About Model Context Protocol (MCP)
What Model Context Protocol (MCP) is
MCP is an open standard from Anthropic for connecting AI tools to data sources, knowledge layers, and external services. By implementing MCP on the server side, DearTech-OS makes its knowledge graph queryable from any MCP-compatible client without per-tool integrations.
MCP support
DearTech-OS is an MCP server. The graph, search, ingest, and audit tools are exposed through the protocol and available to any MCP-compatible client.
Setup
Connect DearTech-OS to Model Context Protocol (MCP)
Configuration takes a few minutes. The exact paths depend on your client version, but the shape of the setup is consistent.
Identify your MCP client
Confirm your AI tool supports MCP. Most modern AI clients (Claude Desktop, Claude Code, Cursor, ChatGPT, Codex) support MCP natively, with more adopting it regularly.
Configure the MCP server endpoint
Add mcp.deartech.co as an MCP server in your client's MCP configuration. The exact location of that config depends on the client (claude_desktop_config.json, ~/.cursor/mcp.json, etc.).
Authenticate
Use the credentials from your DearTech-OS deployment. SSO and role-based access (viewer, maintainer, admin) control what the connected client can see and do.
Use the tools
Once connected, the client has access to graph search, read, query, ingest, and audit tools. The client can now answer questions grounded in your company graph.
Use cases
What you can do with DearTech-OS in Model Context Protocol (MCP)
Multi-AI-tool teams
If your team uses Claude, ChatGPT, and Cursor, all three connect to the same DearTech-OS graph through MCP. Knowledge stays consistent across tools.
Custom agents
Build internal AI agents on the MCP SDK and connect them to DearTech-OS. The same graph serves chat clients, IDEs, and custom workflows.
Future-proof against AI vendor change
If a new AI tool launches with MCP support, it connects to DearTech-OS the same day. Your knowledge layer is decoupled from any single AI vendor.
Vendor-neutral knowledge infrastructure
DearTech-OS is the protocol-level layer. AI tools come and go; the company graph stays.
FAQ
Common questions
What is MCP?
Model Context Protocol is an open standard from Anthropic for connecting AI tools to external data sources, knowledge layers, and tools. It standardizes the interface so AI clients and external services do not need per-tool integrations.
Which AI tools support MCP today?
Claude Desktop and Claude Code support MCP natively. Cursor supports MCP. ChatGPT and Codex support MCP. The list grows as more AI vendors adopt the protocol. Any MCP-compatible client can connect to DearTech-OS.
Do I need to build a custom integration per AI tool?
No. That is the point of MCP. DearTech-OS is one MCP server. Every MCP-compatible AI tool gets access through the same protocol. Build once, connect everywhere.
Can I build my own MCP client to connect to DearTech-OS?
Yes. The MCP SDK is open. Custom agents, internal tools, and one-off scripts can all connect to mcp.deartech.co using the standard MCP client libraries.
Other integrations
Connect DearTech-OS to your other AI tools
DearTech-OS is AI-tool-agnostic. The same knowledge graph serves whichever MCP-compatible client your team uses.
Want DearTech-OS in your Model Context Protocol (MCP) sessions?
In a 30-minute call, we'll map where your company knowledge is scattered today and identify the first DearTech-OS layer worth building.