Modem MCP Server gives AI agents a unified view of customer feedback from Slack, Discord, GitHub, Zendesk, Intercom, email, and call transcripts, letting them search customer issues, feature requests, and complaints, analyze affected users and companies, update topics, and trigger actions through Modem’s agent.
Encrypted at rest, isolated from the model
Resolved from an AES-256-GCM vault at the moment of the call and attached to the request — the model never sees the secrets.
Try asking
Start a durable Modem agent conversation. Returns immediately with conversation_id and run_id; poll modem_agent_get_run for the result. Before your first call, read the skill: `modem_skills({ name: "agent-runs" })`.
Get the current status and terminal result of a durable Modem agent run.
Start a durable follow-up turn in an existing MCP conversation.
Cooperatively cancel one exact durable Modem agent run. Completed external actions are not reversed.
Skills that teach you how to use THIS Modem MCP server's tools. Call before your first `search_modem`, `modem_agent_invoke`, or write-tool call, or when unsure which Modem tool to use. Pass `name` for the full skill, one of `search-modem`, `agent-runs`, `write-tools`. Omit `name` to list them with descriptions. This is not the Skills page in dashboard settings.
Answer a natural-language question about your Modem data and get matching rows back. Read-only. It does not run the full Modem agent or spend agent credits. Send a `question`, or send only the `cursor` from a previous result to fetch its next page. Topic `priority` in the rows is an integer: very_low (-100), low (-50), default (0), high (50), very_high (100). Before your first call, read the skill: `modem_skills({ name: "search-modem" })`.
Search or read Modem's public docs to answer how Modem works: setup, integrations, this MCP server and its scopes, the API, and automations. Not for your organization's data; use `search_modem` for that. Send a `query` to search, or a `page` path from a search result (such as `api/modem-mcp-server`) to read the whole page. Send exactly one. Read-only, spends no agent credits.
Update a single topic (priority, lifecycle state, archived status, issue type, or keywords). Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Update multiple topics at once (e.g. archive or re-prioritize a set of topics). Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Merge one or more source topics into a target topic. Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Create one or more companies in the organization. Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Update fields on one or more existing companies. Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Merge one or more source companies into a target company. Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Associate one or more people with a company. Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Update fields on one or more existing people. Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
Merge one or more source people into a target person. Before your first call, read the skill: `modem_skills({ name: "write-tools" })`.
One endpoint, the same key, whichever client you use.
~/Library/Application Support/Claude/claude_desktop_config.json (Mac) · %APPDATA%\Claude\claude_desktop_config.json (Windows)
Replace API_KEY with your own key.
Already have an "mcpServers" section in your config? Just add the server entry inside it.
Discovery, routing, credentials, tool scoping and execution logs all happen at the gateway→connections stay ACTIVE with no work from you
Modem MCP runs through a gateway that holds the credentials, scopes the access and records every call.
Managed auth, hosted MCP servers, and every Gmail tool your agent needs.
Free to start.