Integration

MCP — Model Context Protocol

Connect Claude Desktop, Cursor, and any MCP-compatible client to your ZainaBot projects. Ask Claude about your open commitments, query your project history, list your action items — all from your own AI client.

1. Generate an API key

API keys are tied to your account and used to authenticate every MCP request. Pro+ required.

Open API keys page →

2. Configure Claude Desktop

Open ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows) and add:

{
  "mcpServers": {
    "zainabot": {
      "type": "http",
      "url": "https://mcp.zainabot.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY_HERE"
      }
    }
  }
}

Restart Claude Desktop after editing. The MCP server icon should appear in the chat input.

3. Try it

In Claude Desktop, try asking:

"What action items do I have open this week?"
"List my ZainaBot projects."
"Summarize what I know about Acme from my project memory."
"Ask Yusuf how to debug a slow Postgres query." (uses chat_with_coach)
"Add a follow-up: email Acme by Friday." (uses create_action_item)
"Start a sales-call practice with Faris." (uses start_practice_session)
"Run the blog-to-thread workflow on: why pricing pages lie." (uses run_workflow)
"How much of my monthly token quota have I used?" (uses get_usage)

Claude picks the right ZainaBot tool automatically based on the question.

Available tools

Read

list_coaches — browse the 50+ AI coaches, optionally by domain.
list_projects — your active projects.
list_action_items — open commitments across projects.
query_project_memory — natural-language search over a project's sessions + docs.

Write

chat_with_coach — single-turn chat with any coach. Continues an existing session if you pass one.
create_action_item — add a new commitment, optionally tied to a project.
update_action_item — mark done, dismissed, or reopen.
delete_action_item — permanently remove an item.
start_practice_session — start a roleplay. Returns a session_id; reply via chat_with_coach.
attach_session_to_project — add an existing chat / practice / live session to a project.

Coming soon: run_workflow (workflow execution + run status).

Rate limits + quota

Per-API-key buckets, 1-minute windows. Exceeding a bucket returns an error mentioning the bucket + retry-after.

Read (list_*) — 60/min
Write (create / update / delete / attach / start practice) — 60/min
Chat (chat_with_coach) — 30/min
Query (query_project_memory) — 20/min

Chat + query calls also count against your monthly token quota (Pro 1M, Pro+ 2M, Enterprise 10M).

Server details

Transport: HTTP (JSON-RPC 2.0)
URL: https://mcp.zainabot.ai/mcp
Auth: Authorization: Bearer zb_pk_*
Protocol: 2024-11-05

Learn more about MCP: modelcontextprotocol.io

Questions? Open a support ticket →