VORIM
We use cookies

We use cookies to analyze site traffic and improve your experience. You can choose to accept all cookies or only essential ones. See our Privacy Policy.

MCP server

Nineteen tools over stdio for Claude Desktop, Cursor, VS Code, Goose and any other MCP client. Register, check, emit and verify without leaving the editor.

MCP Server

The Vorim MCP server exposes all Vorim operations as tools for Claude Desktop, Cursor, VS Code, Google Antigravity, and any MCP-compatible AI client. Agents can register themselves, check permissions, log actions, and verify trust — all through natural language.

npm · GitHub · Smithery

Installation

npm install -g @vorim/mcp-server
bash

Configuration

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "vorim": {
      "command": "npx",
      "args": ["@vorim/mcp-server"],
      "env": {
        "VORIM_API_KEY": "agid_sk_live_..."
      }
    }
  }
}
typescript

For Cursor, add to .cursor/mcp.json in your project root with the same format.

Berd & Goose (Block)

Berd is Block's desktop app for AI agents, running on the Goose engine — an MCP client. Add Vorim as a Standard IO extension and every agent you run in Berd (or Goose) gets identity, permission checks before actions, and a signed audit trail.

Berd / Goose Desktop: open the sidebar → ExtensionsAdd custom extension. Set Type to Standard IO, Command to npx @vorim/mcp-server, and add an environment variable VORIM_API_KEY = your key.

Goose CLI: run goose configureAdd ExtensionCommand-Line Extension, then enter the command npx @vorim/mcp-server.

Or add it directly to Goose's ~/.config/goose/config.yaml:

extensions:
  vorim:
    enabled: true
    type: stdio
    cmd: npx
    args:
      - "@vorim/mcp-server"
    env_keys:
      - VORIM_API_KEY
    timeout: 300
yaml

Store the key with goose configure so it lands in Goose's secret store; env_keys tells Goose which secret to pass to the server at launch.

Cosine

Cosine ships autonomous coding agents (Lumen) that plan a change, write it, and open a pull request without a human in the loop. Its CLI is an MCP client, so adding Vorim gives every agent Cosine runs a cryptographic identity, a permission check before it acts, and a signed audit trail. When an agent acts unsupervised, that signed record of what it touched and under whose authority is what you hand an auditor afterward.

Fastest: add Vorim from the CLI:

cos mcp add --transport stdio -e VORIM_API_KEY=agid_sk_... vorim -- npx @vorim/mcp-server
bash

Or edit ~/.cosine/mcp.json directly:

{
  "mcpServers": {
    "vorim": {
      "transport": "stdio",
      "command": "npx",
      "args": ["@vorim/mcp-server"],
      "env": {
        "VORIM_API_KEY": "agid_sk_..."
      }
    }
  }
}
json

Press Ctrl+J in the Cosine TUI and pick MCP servers to check that Vorim connected. On Windows the config lives at %USERPROFILE%\\.cosine\\mcp.json.

Available Tools (19)

ToolDescription
vorim_pingCheck API health and connectivity
vorim_register_agentRegister a new agent with a cryptographic identity
vorim_get_agentGet agent details by ID
vorim_list_agentsList all agents with pagination and filtering
vorim_update_agentUpdate agent metadata
vorim_revoke_agentPermanently revoke an agent
vorim_check_permissionCheck if agent has a permission scope (<5ms)
vorim_grant_permissionGrant a permission scope with optional expiry/rate limits
vorim_list_permissionsList all active permissions for an agent
vorim_revoke_permissionRevoke a specific permission scope
vorim_emit_eventLog an audit event for an agent action
vorim_export_auditExport signed audit bundle with SHA-256 manifest
vorim_verify_trustVerify agent trust score (public, no auth)
vorim_register_ephemeralRegister an ephemeral did:key agent that auto-expires; returns agent_id, did:key, and keypair
vorim_delegate_credentialDelegate a scoped credential to an agent, with optional rate limits and expiry
vorim_request_tokenRequest a short-lived, scoped access token for an agent and provider
vorim_list_delegationsList credential delegations, optionally filtered by agent
vorim_onboard_startStart device-authorization onboarding for a user with no API key; returns a user code and activation URL
vorim_onboard_checkCheck whether the user approved onboarding and retrieve the issued API key

Example Usage

Once configured, you can use natural language in Claude, Cursor, or any MCP client:

// In Claude Desktop or Cursor, just ask:
"Register an agent called invoice-processor with read and execute permissions"
"Check if agent agid_acme_a1b2 has permission to execute"
"Log a tool_call event for agent agid_acme_a1b2: action=process_invoice, result=success"
"What's the trust score for agent agid_acme_a1b2?"
"Export the audit trail for the last 7 days"
typescript

Agent Discovery

Vorim AI publishes a machine-readable Agent Card for automated discovery:

curl https://vorim.ai/.well-known/agent.json
bash

This follows the A2A (Agent-to-Agent) discovery pattern, allowing other agents to programmatically discover Vorim's capabilities, endpoints, and authentication.

Need Help?

Book a demo for a walkthrough, or contact us for support. For enterprise needs, reach out at sales@vorim.ai.