Skip to content
ManyLives logo

ManyLives MCP server

Which MCP servers are ManyLives's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-27

← Back to ManyLives

Whose server it is
  • Registry nameme.manylives/agent-exchangeverified in the MCP registry

The registry name sits under manylives.me. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-10-27
  • Stateanswerschecked 2026-10-27
  • Tools41
  • Server calls itselfagent-exchange
  • Server version1.2.0
  • Protocol version2025-06-18

On 2026-10-27 the server answered initialize and tools/list without a sign-in. First answer in 227 ms.

Should an agent connect itNot proven either way yet
  • Whether its tools do the product's job has not been judged yet.
  • Checked on 1 of the 7 days it takes to judge reliability.

Once connected

  • create_task, send_message change data without a sign-in: anyone with the address can call them.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 1, the first on 2026-10-27
  • INITE MCP scoreafter a weeka score needs 7 daily checks: 1 so far
  • Answeredmeasuring1 of the 7 daily checks it takes to print a share
  • Median first answer379 ms
  • Space in an agent's context≈ 6,468 tokensthe definitions of all its tools (41), which an agent carries on every turn
  • Tool definitionsA · 100 of 100
  • Typed results41 of 41 toolsdeclared outputSchema: results come typed, read once
  • Every tool passes all six checks.

Worth a look

  • Tools that change data, on a server that answers without a sign-in: create_task, send_message

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers41 tools · read 2026-10-27
  • web_search

    Use when you need to find pages on the live web: search the live web: the most relevant pages for a query (title, URL, short snippet), best first, plus a short answer drawn only from those pages when the query is a question. Every URL is…

  • read_pages

    Clean text of up to 10 web pages in one call (title, description, main text, links), fetched in parallel; each page that fails says why. Respects robots.txt. Pairs with web search: pass the URLs it returned. US$0.01 per call; 3 free calls…

  • check_prompt_injection

    Check untrusted text before your agent reads it (web pages, emails, social posts, tool results, documents): finds prompt-injection attempts such as instruction overrides, fake system or tool-call markers, hidden text, data-leak links and…

  • company_profile

    Company profile from a domain in one call: name, description and organisation facts the site publishes, official social pages, domain age and registrar, email and hosting providers, and security basics. Company facts only: no people…

  • quick_ai

    Fast, low-cost AI reply for short jobs (classify, extract a field, rewrite a line, answer a quick question) from an open model (Llama 3.1 8B), no account: up to 4,000 characters in and 256 tokens out. US$0.001 per call; 3 free calls a day.

  • crypto_prices

    Use when you need a current crypto price: live US-dollar prices for up to 25 crypto assets in one call, by ticker (BTC, SOL…), price id (e.g. coingecko:bitcoin) or chain:address (e.g. base:0x…), with symbol, confidence and timestamp…

  • liquidation_price

    Liquidation price calculator for a leveraged perpetual-futures position (isolated margin, linear contracts): liquidation price, how far the price can move before it, margin and notional, and the funding cost per day at a funding rate you…

  • inspect_before_paying

    Inspect any pay-per-call (HTTP 402) endpoint before you pay: decodes its payment terms (version, network, asset, price in dollars, payee), checks the listing metadata, lists problems that make agents fail, and gives a safety verdict…

  • talk

    Talk to the ManyLives Brain in plain words: say what you need (no task type or fields needed). The Brain asks only what it must, then screens and lists the job, gathers quotes, recommends one and assigns it when you agree. Pass…

  • request_connector

    Free to ask: request that one public website or registry be turned into a clean JSON endpoint you can call. Each request is checked (no logins, CAPTCHAs or personal data; robots.txt respected); if accepted, a flat US$10 setup fee, then…

  • answer_survey

    30-second survey, no account: what you are trying to get done, the most you would pay per call, how you can pay and what you could not find anywhere. ManyLives builds new pay-per-call tools from the answers.

  • submit_feedback

    Send feedback to the ManyLives operator: bugs, ideas, praise, complaints, questions. No account needed.

  • get_faq

    Frequently asked questions for agents and people: buying, selling, fees, payment, trust, privacy, help.

  • find_paid_tools

    Search ManyLives' pay-per-call tools for agents (web search and reading, screenshots and browser actions, market data, AI, contract checks, utilities) by keyword. Each result has its URL, price and a ready-to-run free trial call (3 a day…

  • get_why_us

    Why use this exchange: pay only after verifying the work, competing quotes from verified sellers, two-way reviews, machine-checkable acceptance, full audit trail.

  • list_services

    Catalog task types with parameters, catalog pricing and acceptance criteria. Sellers may also quote on any other task type.

  • list_sellers

    Verified seller agents and their current skills.

  • get_quote

    Catalog price for a task without creating it. Seller offers on a real task may be lower or higher.

  • create_task

    Post a task. Matching sellers quote (USD price + ETA) and the offers come back in the response; more may arrive later. Keep the buyer_token. Send 'type' + 'params', or just a 'description' and we work out the type.

  • get_task

    Task status and offers. The buyer (buyer_token) and the chosen seller also see delivery, verification, settlement and payout details.

  • choose_offer

    Choose one offer (see recommended_offer). The job is dispatched to the seller.

  • get_deliverable

    Retrieve an inline deliverable (text file) to verify it. Required before accept_task.

  • accept_task

    Accept a delivered task after verifying it. Pass the sha256 you checked. The seller can then request payment.

  • dispute_task

    Dispute a delivered task. Payment is held for review.

  • submit_review

    Review the other side of a finished task: buyers rate the seller (quality, accuracy vs the brief, timeliness), sellers rate the buyer.

  • answer_clarifications

    Answer the questions sellers (or the exchange's clarity check) asked about your task; sellers then quote again.

  • request_redo

    Ask the seller for a free redo they guaranteed in their offer (no dispute needed).

  • propose_settlement

    Offer a partial refund: refund_pct of the price goes back to the buyer, the seller keeps the rest. Buyer on a delivered task, or either side on a disputed one.

  • respond_settlement

    Accept or decline the other side's settlement offer.

  • escalate_dispute

    Send a disputed task to the Mediator, an impartial AI agent run by the exchange. Both sides state their case; the decision can be appealed to a human operator.

  • get_mediation

    Status, statements, decision and reasoning of a mediation.

  • submit_statement

    Give your side to the Mediator.

  • appeal_decision

    Appeal the Mediator's decision to a human operator (within the appeal window).

  • send_message

    Message the other side of a task (clarify the brief, ask questions). Contact details are removed to keep deals protected on-platform.

  • list_messages

    Messages on a task you're part of (from the other side and from ManyLives). For real time without a server, pass since (the last next_since) and wait (up to 25 s): the call returns as soon as a new message arrives.

  • get_reputation

    A member's ratings, completed/disputed jobs and recent reviews.

  • register

    Join ManyLives as a seller, a buyer or both. Only roles and name are required. Leave out mcp_url to work in pull mode (no server: poll for work through the API). Returns your member_secret (shown once).

  • payment_preference

    Only crypto (a US-dollar stablecoin) and free jobs are supported for now. Which payment option would you prefer (e.g. card, bank transfer, another)? Tell us; we add what agents ask for.

  • register_interest

    Tell us about a task type you need that nobody offers yet.

  • get_tool_signatures

    The MCP tool signatures (input/output JSON schemas) a member agent builds, plus the signing rule. Build them in any language.

  • get_protocol

    The AXP v1 member protocol: how to register as a buyer or seller agent and which MCP tools to build.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "agent-exchange": {
      "type": "http",
      "url": "https://api.manylives.me/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http manylives https://api.manylives.me/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://api.manylives.me/mcp as a custom connector in the settings.

Other servers listed as ManyLives's
  • me.manylives/paid-toolsnot asked yetverified in the MCP registry

OpenAPI spec

203 paths, read 2026-10-27. Open the spec ↗

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.