MCP setup

Connect dogo to your agent.

dogo exposes its full exec-search loop as an MCP server over Streamable HTTP. Once connected, your agent can run an entire search — from brief to drafted outreach — without touching the UI.

01

Create an account

Sign up at dogoapp.ai.

02

Generate an API key

In the dogo app, go to Settings → API keys and create a new key. The full key is shown once — copy it somewhere safe. It looks like:

dogo_sk_AbC3kZ9xYw...
03

Add the config

Add one block to your agent's MCP configuration, replacing dogo_sk_... with your key.

Claude.ai (web)

In Claude.ai, go to Settings → Connectors → Add custom connector and fill in:

  • Name: Dogo MCP
  • URL: https://dogoapp.ai/api/mcp
  • OAuth Client ID: any value (e.g. dogo)
  • OAuth Client Secret: paste your dogo_sk_... key

Claude Desktop

Open ~/.claude.json and add:

{
  "mcpServers": {
    "dogo": {
      "type": "http",
      "url": "https://dogoapp.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer dogo_sk_..."
      }
    }
  }
}

Codex / other MCP clients

Use the same URL and Bearer token. Any client that supports the Streamable HTTP MCP transport (the 2025 standard) will work.

Available tools

What the agent can do

Once connected, your agent has access to the full exec-search loop as 16 purpose-built tools — not a CRUD layer over the database.

  • create_missionCreate a mission
  • draft_mission_planTurn a brief into weighted criteria
  • run_sourcingSource and score candidates
  • get_ranked_candidatesRead the ranked longlist
  • shortlist_candidatesAdd candidates to the shortlist
  • enrich_contactsFind verified email and phone
  • generate_outreachDraft personalized first-touch messages
  • add_to_poolAdd a person to the talent pool
  • search_talent_poolSearch the firm's talent pool
  • researchResearch a topic with web citations
  • expand_similarFind more candidates like a specific one
  • expand_websearchRun a deeper web search pass

Full schema and request/response examples in the API docs.