MCP.Market

Txn MCP server

by txn.dev·dev.txn/mcp-server·v0.2.2
No reviews yet

Wallets, payments and human hiring for AI agents: pay other agents, escrow a budget, hire a person.

B78/100grade B
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
B78/100

full report

Adoption
New

0 stars28 downloads/wk

Reviews

Write one

Nobody has reviewed Txn yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

Txn tools (13, 3 write)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

  • assign_job

    Choose which applicant does a job you posted. Everyone else who applied is told no. If they asked for more than the posted budget, pass agreed_amount to accept their price and the extra is escrowed too.

  • create_walletwrite action

    Create a new agent wallet for sending and receiving payments

  • dispute_job

    Reject delivered work and say what is wrong with it. The person can then fix it and resubmit. The money stays in escrow: it is neither paid out nor returned to you, because delivered work cannot be taken back for free.

  • find_humans

    Find real people available for hire for work an AI cannot finish alone: anything needing physical presence, a legal signature, local knowledge, licensed expertise, a human judgement call, or an account only a person can open. Returns each person's skills, rate floor, rating, and completed job count.

  • fund_wallet

    Generate a Stripe checkout link to add funds to an agent wallet. Returns a URL the user can open to complete payment.

  • get_balance

    Check the current balance of a specific agent wallet

  • get_job

    Check on a job you posted or a person you hired: its status, who applied, what they delivered, and how much is still held in escrow. Call this before releasing payment so you can read the delivered work.

  • hire_human

    Hire a specific person to do a piece of work. The budget is moved into escrow immediately and is only paid out when you release it, so the money is committed but not spent. Find someone with find_humans first. The person is told the work was commissioned by an AI agent.

  • list_jobs

    List the jobs you have posted, newest first. Filter by status to find work waiting on you: 'delivered' is work you need to review and pay for.

  • list_wallets

    List all agent wallets in your organization

  • paywrite action

    Transfer funds from one agent wallet to another. Use when an agent needs to pay for a completed task or service.

  • post_jobwrite action

    Post a job to the public board and let people apply, instead of picking someone yourself. The budget is escrowed when you post. Use this when you do not know who should do the work; use hire_human when you already do. Review applicants with get_job, then choose one with assign_job.

  • release_payment

    Accept delivered work and pay the person. This releases the escrow and cannot be undone, so read the deliverable with get_job first. If the work is not right, use dispute_job instead and they get a chance to fix it.

Public scan report

scanner v0.1.10 · 2026-10-16 · same rubric, same numbers if you re-run it

no findings
  • Code scan2 source files scanned25/25
  • –Live reliabilityno gateway calls yet and no remote to proben/a
  • –Tool poisoningtools not inspected (local package is not executed); not countedn/a
  • Auth qualitystatic API keys via environment variables6/15
  • Maintenancelast push 44 days ago12/15
  • Maintainer identitynamespace and repository owner differ; GitHub account older than a year; website matches verified namespace8/10
Overall 78/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

How Txn compares

Against every graded MCP server, from our own scans and uptime checks. Updated after each scan run.

Safety score
78/100
higher than 57% of 48,243 servers
Tools
13
the typical server has 7
Tools that change things
23%
typical: 0%. Fewer write tools means less an agent can break
Downloads a week
28
more than 15% of 12,306 servers with a package

What its tools let an agent do: 9 read · 1 create · 2 change · 1 send or post

Install Txn in Claude Code, Cursor or VS Code

Runs npx -y @txn-dev/mcp-server on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add mcp-server -- npx -y @txn-dev/mcp-server
Add to Cursor

What the publisher says

From the Txn repository's README, as published. We do not edit it. Read it on GitHub

@txn-dev/mcp-server

MCP server for txn.dev - let any AI assistant manage agent payments.

Works with Claude Desktop, Cursor, Windsurf, and any MCP-compatible client.

Setup

Add to your MCP client config:

{
  "mcpServers": {
    "txn": {
      "command": "npx",
      "args": ["-y", "@txn-dev/mcp-server"],
      "env": {
        "TXN_API_KEY": "txn_live_..."
      }
    }
  }
}

Get your API key at txn.dev/dashboard/api-keys. A txntest key moves sandbox money through the same tools, so try it there first.

Claude Code, in one line:

claude mcp add txn -e TXN_API_KEY=txn_test_... -- npx -y @txn-dev/mcp-server

Tools

What you can do

Once connected, ask your AI assistant things like:

  • "Create a wallet called research-agent"
  • "How much is in the translator wallet?"
  • "Pay $0.50 from coordinator to summarizer for the report"
  • "List all my agent wallets and their balances"
  • "Fund the coordinator wallet with $20"

Environment Variables

License

MIT

Nothing above is checked by us. What we check is on the safety report.

Txn: common questions

Is Txn MCP server safe?
Mostly: it is graded B (78/100). Read the Txn safety report
How do I install Txn?
It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does Txn need an API key?
Yes. The registry entry asks for TXN_API_KEY.
Is Txn maintained?
The last commit was 64 days ago (2026-09-03). The latest release is v0.2.2.
How does Txn compare with other MCP servers?
safety score 78/100 (higher than 57% of 48,243 servers); tools 13 (the typical server has 7); tools that change things 23% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.