Skip to content
MCP server

Connexagents MCP

Fallback for stuck agents: hand a blocked call to a specialist and pay only on success

Runs locallyStatus and history
Install
Pricing
Paid
Platforms
CLI, API
Runs
Locally
Version
0.2.2
Listing updated
3 days ago

About Connexagents MCP

Connexagents MCP is a paid MCP server for developer tools. It runs locally on your machine and is started by your AI client.

The first problem it targets is API authentication. When an agent hits a 401 or 403 it cannot get past, it calls find_specialist with a summary of the gap, acceptance criteria, a deadline and an optional domain hint. ConnexAgents proposes a specialist, nothing goes out without human confirmation, and the specialist gets temporary, scoped access.

Success is checked mechanically: the failing call is replayed and must return a 2xx, timestamped, and the record is published whether it passes or fails. A handoff id can be polled, and a separate tool lists match proposals for a linked agent.

The price is 99 euros excluding VAT per proven unblock and nothing otherwise, with the first unblock free and no subscription. Registration takes about five minutes. You need a ConnexAgents agent key, and the server runs over stdio with the Claude Code, Cursor or Claude Desktop clients.

Install

Run in a terminal

claude mcp add connexagents-mcp --env CONNEXAGENTS_TOKEN='<CONNEXAGENTS_TOKEN>' -- npx -y connexagents-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "connexagents-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "connexagents-mcp"
      ],
      "env": {
        "CONNEXAGENTS_TOKEN": "<CONNEXAGENTS_TOKEN>"
      }
    }
  }
}

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "connexagents-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "connexagents-mcp"
      ],
      "env": {
        "CONNEXAGENTS_TOKEN": "<CONNEXAGENTS_TOKEN>"
      }
    }
  }
}

Add to .vscode/mcp.json

{
  "servers": {
    "connexagents-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "connexagents-mcp"
      ],
      "env": {
        "CONNEXAGENTS_TOKEN": "<CONNEXAGENTS_TOKEN>"
      }
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "connexagents-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "connexagents-mcp"
      ],
      "env": {
        "CONNEXAGENTS_TOKEN": "<CONNEXAGENTS_TOKEN>"
      }
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "connexagents-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "connexagents-mcp"
      ],
      "env": {
        "CONNEXAGENTS_TOKEN": "<CONNEXAGENTS_TOKEN>"
      }
    }
  }
}

Configuration

  • CONNEXAGENTS_API_URLAPI URL, ending in /api/v1
  • CONNEXAGENTS_TOKENConnexAgents agent key (cxa_...)Required
  • CONNEXAGENTS_AGENT_IDUUID of the agent linked to this key. Required only for list_match_proposals.

Tools this server provides (3)

Taken from the project documentation; the running server may differ.

find_specialist
Reports a block and requests a specialist, with gap summary, acceptance criteria and deadline.
get_handoff_status
Returns the status of a handoff.
list_match_proposals
Lists specialist proposals for the linked agent.

What it does

Specialist matching
Proposes a specialist for the failing call, with human confirmation.
Mechanical proof
Replays the failing call and records an HTTP 2xx result.
Public record
Publishes successes and failures.
Scoped access
Specialists receive temporary, limited access.
Handoff status
Check progress of a delegation by id.
Pay on success
Billing only occurs when the agent is unblocked.

Good for

  • Get an agent past an OAuth2 plus HMAC authentication error
  • Delegate a stuck API integration to a specialist
  • Audit how a blocker was resolved from the public record

Pricing

Paid

EUR 99 excl. VAT per proven unblock; first unblock free

Prices change. Check connexagents.com before you buy.

Connexagents MCP FAQ

What does an unblock cost?

99 euros excluding VAT, charged only if the agent is unblocked; the first is free.

How is success decided?

The failing call is replayed and must return HTTP 2xx.

Does anything happen without my approval?

No. The site says nothing goes out without human confirmation.

What credentials are needed?

A ConnexAgents agent key in CONNEXAGENTS_TOKEN.

Is Connexagents MCP free?

No. Connexagents MCP is a paid product.

How do I add Connexagents MCP to Claude Code?

Run this command in your terminal: claude mcp add connexagents-mcp --env CONNEXAGENTS_TOKEN='<CONNEXAGENTS_TOKEN>' -- npx -y connexagents-mcp

What are the best alternatives to Connexagents MCP?

The closest alternatives to Connexagents MCP are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to Connexagents MCP

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ouroboros

    MCP server

    Spec-first layer that pins requirements and verifies results of AI coding agent work

    Runs locally
    • Open source
    • 6.4K GitHub stars
    • Developer tools