Skip to content
Amazap logo

Amazap MCP server

Which MCP servers are Amazap'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 Amazap

Whose server it is
  • Registry namecom.amazap/amazapverified in the MCP registry

The registry name sits under amazap.com. 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
  • Tools24
  • Server calls itselfamazap
  • Server version0.5.21
  • Protocol version2025-06-18

On 2026-10-27 the server answered initialize and tools/list without a sign-in. First answer in 254 ms.

Should an agent connect itNot proven either way yet
  • Its tools do every use case we checked: read web pages for fewer AI tokens; squeeze long CI failure logs before analysis.
  • Checked on 1 of the 7 days it takes to judge reliability.

Once connected

  • For "read web pages for fewer AI tokens", the likely first call is read_webpage.
  • For "squeeze long CI failure logs before analysis", the likely first call is squeeze_log.

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
  • What an agent can do with it2 of 2 use cases
  • Answeredmeasuring1 of the 7 daily checks it takes to print a share
  • Median first answer280 ms
  • Space in an agent's context≈ 6,269 tokensthe definitions of all its tools (24), which an agent carries on every turn
  • Tool definitionsA · 100 of 100
  • Typed results24 of 24 toolsdeclared outputSchema: results come typed, read once
  • Every tool passes all six checks.

What an agent can do with it

  • can do · read web pages for fewer AI tokens
  • can do · squeeze long CI failure logs before analysis
  • not sure · extract plan prices from pricing pages

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

None of the five warning signs in these definitions.

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 offers24 tools · read 2026-10-27
  • read_webpage

    Clean article Markdown. Use it when the page is long or cluttered. Skip it on a short or already clean page: the quote compares the tokens both ways and does not charge for that check. Savings are compared with the page's visible text…

  • lightning_node

    Lightning node snapshot by pubkey. Pass pubkeys (up to 10) for one call.

  • lightning_liquidity

    Current Lightning inbound-liquidity prices.

  • read_feed

    RSS/Atom/JSON feed as compact items. Saves 27-81%; on lean feeds skip unless you cannot parse XML.

  • make_qr

    Make a PNG QR code for an invoice, URL, or text. Put the QR URL in your own message.

  • squeeze_log

    Shrink a failed CI log. Pass a GitHub Actions job link or the pasted log.

  • page_metadata

    Title, description, OG/Twitter, and JSON-LD for a URL. Savings are compared with the page's visible text, not its HTML. Not for reading the page.

  • read_pdf

    PDF to Markdown, pages e.g. 1-3 (default first 30). Sponsor pages are omitted and listed in adPages. Use only if you cannot run pdftotext: output is about pdftotext size, no token saving.

  • repo_digest

    GitHub repo overview (tree, README, entry points, commits), ~750-1.5k tokens. Use on medium/large repos. Skip if you only need the README: read it directly.

  • search_catalog

    Find paid Amazap tasks by keyword. Results: price, value (est. saving), buy call.

  • describe_sku

    Parameter schema, schema confidence, and exampleCall for one SKU. total_cost is the all-in price only. Default USD. If the record is untested and totalSats is over 500, tell the user before buying. When Lightning payments are temporarily…

  • quote

    Free price check incl. value estimate (raw vs output tokens, break-even $/M). Call only to change quantity/currency or check value before buying. When Lightning payments are temporarily paused, the result includes paymentsPaused and…

  • buy

    Spends prepaid Amazap service credits on a third-party API call, within the account's spend limits. A purchase can't be undone. Only failed calls are re-credited, up to 1,000 sats a day and 500 sats a day per provider. Direct purchases…

  • buy_best

    Spends prepaid Amazap service credits on a third-party API call, within the account's spend limits. A purchase can't be undone. Only failed calls are re-credited, up to 1,000 sats a day and 500 sats a day per provider. Direct purchases…

  • orders

    Recent buys, one short row each: order id, time, status, charged sats, refunded sats, and net sats. Charged is the gross debit. Net is charged minus refunded. Pass include_results for each stored result, or get_order for one order's notes…

  • get_order

    One order, including the stored result. A seller payment still in flight is checked again. Does not start a new payment.

  • claim_failed_call

    Ask for a service-credit review of one paid call. For a basket, pass line or sku. Claims older than 72 hours are rejected. This does not move sats or cash.

  • balance

    Remaining sats and the hourly buy limit. maxBuysPerHour and remainingBuysHour are on this result. Do not call this before buying. A paid top-up is credited from the payment event. session_budget_sats sets the session cap up to the saved…

  • spend_summary

    Today's spend, remaining budgets, and the last 10 buys. Charged is the gross debit. Refunded is separate. Net is charged minus refunded. Read-only. Does not spend sats.

  • top_up

    Only creates an invoice for the human to pay. It does not move sats and does not spend the credit balance. Put the QR image and the QR URL in your own message. Pay from any Lightning app. Do not paste the invoice. Pass include_invoice…

  • catalog_stats

    Read-only public listing counts by protocol. Does not spend sats.

  • report_outcome

    Report how a paid call went. Spends nothing. A service credit may be added to the balance when the paid call was at least 20 sats. A smaller call is recorded and is not credited. The credit is not cash. Report credits pause after a few…

  • report_sku

    Flag a bad schema, a junk or incorrect response, or a dead SKU. Sends a short reason and spends nothing. Limited per agent per day.

  • list_my_endpoint

    Probe one public https endpoint for an L402 challenge and a prefilled listing draft. Does not spend sats and does not publish the listing.

How to connect

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

{
  "mcpServers": {
    "amazap": {
      "type": "http",
      "url": "https://amazap.com/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http amazap https://amazap.com/mcp

Add to CursorAdd to VS Code

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

OpenAPI spec

18 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.