Brain MCP Server: workspace memory as tools

The Brain MCP server exposes your Alfera workspace brain as callable tools in any MCP-compatible agent. Point Claude Code, Cursor, or your own harness at a single URL and your agent can search memory, read context, teach new facts, and ask the Alfera worker for answers. The server is streamable HTTP and stateless: no session to open, resume, or close. Every request carries its own key, so restarting your client or load-balancing across instances changes nothing.

Endpoint

https://api.alfera.ai/v1/mcp

Key facts

Streamable HTTP, stateless. No session to open or close. Every request is independent.
Protocol versions. The server speaks MCP 2025-06-18 and accepts 2025-03-26 and 2024-11-05.
Same brain, same keys. It uses the same workspace brain and the same API keys as the Brain API. Disclosure rules are identical: an external key receives only profile_fact, preference, decision, and observation, and never facts about your team, agent_inferred facts, or facts about an end user the call did not name.

Connect

Create a key in the Alfera app under Settings → API Keys, pick its scopes and audience, and export the secret:

export ALFERA_API_KEY=ak_live_...

Claude Code
Cursor
Other

Add the MCP server with the claude mcp add command:

claude mcp add --transport http alfera-brain https://api.alfera.ai/v1/mcp \
  --header "Authorization: Bearer $ALFERA_API_KEY"

Add the server to ~/.cursor/mcp.json (global) or .cursor/mcp.json (one project):

{
  "mcpServers": {
    "alfera-brain": {
      "url": "https://api.alfera.ai/v1/mcp",
      "headers": { "Authorization": "Bearer ak_live_..." }
    }
  }
}

Anything that speaks Streamable HTTP can POST JSON-RPC to https://api.alfera.ai/v1/mcp with Authorization: Bearer <key> on every request. X-Api-Key: <key> is accepted as well.The endpoint is POST-only. GET and DELETE return 405 method_not_allowed.

Verify with whoami

Call whoami to confirm the connection. It needs no scope, so it answers for any live key and reports what the connection can do.

curl https://api.alfera.ai/v1/mcp \
  -H "Authorization: Bearer $ALFERA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "whoami",
      "arguments": {}
    }
  }'

string

Human-readable name for this key, for example support-bot.

string

Prefix of the key, for example ak_live_x7f4.

array

Array of scopes this key is allowed to use, for example ["memory.read", "answers.write"].

string

Audience for the key, for example external.

object

Object with requests_per_window and window_seconds, for example { "requests_per_window": 120, "window_seconds": 60 }.

Example response:

{
  "key_name": "support-bot",
  "key_prefix": "ak_live_x7f4",
  "scopes": ["memory.read", "answers.write"],
  "key_audience": "external",
  "rate_limit": { "requests_per_window": 120, "window_seconds": 60 }
}

tools/list only advertises what this key’s scopes allow. A memory.read-only key never sees ask_worker. If a tool is missing from your agent’s list, the key lacks its scope.

Next steps

Browse the full Brain MCP tools reference for arguments, scopes, and usage patterns.
Read the Brain API overview for the REST equivalents and field-level detail.

​Endpoint

​Key facts

​Connect

​Verify with whoami

​Next steps

Endpoint

Key facts

Connect

Verify with whoami

Next steps