Connect an AI agent to Tradeics MCP in minutes. Pick one server first (usually Docs for read-only, or Platform if you already have an API token), prove a tool call, then add more.
If you already use Authentication and REST API V2, this will feel familiar.
Pick a server
| Start with | URL (placeholder) | Why |
|---|---|---|
| Docs MCP | https://mcp.tradeics.example/docs |
No workspace side effects |
| Platform MCP | https://mcp.tradeics.example/platform |
Live data once you have scopes |
| AI MCP | https://mcp.tradeics.example/ai |
Assistive tools after Platform is solid |
Legacy single-path placeholder (transition): https://api.tradeics.example/mcp/ — treat as Platform-equivalent until hosts migrate to per-server URLs.
1-minute setup
- Add a new MCP server in your AI tool.
- Choose transport type: Streamable HTTP.
- Set the URL to your chosen server from the table above.
- Add the authorization header:
Authorization: Bearer <TRADEICS_MCP_TOKEN>- Connect. Confirm tools appear for that server only.
- Run one read tool to prove auth (for example docs search, or list catalogs / categories).
Alternatively, pass ?token=<TRADEICS_MCP_TOKEN> if your host cannot send custom headers. Prefer the Authorization header — query tokens leak more easily through logs. URL-encode tokens when using the query form.
Early access: tool catalogs grow with REST V2 scopes. Missing a tool? Use the REST reference and ask support to enable the right MCP server for your workspace.
Suggested first prompts
Docs MCP
- “Search Tradeics docs for pagination and quote the guide path.”
Platform MCP
- “List vendor sub-categories we can use when syncing catalogs.”
- “Summarize open RFQs that still need supplier responses.”
AI MCP
- “Classify this catalog line into our taxonomy (assistive only).”