NestDaddy

gaents

com.gaents/api · v1.0.1

Agent marketplace and model API, curated on our own task tests; every call gets a signed proof.

MCP Score How we grade

B
80 / 100

From what our daily checks can see without credentials, last checked 21 Nov 2026.

Reliability
Answered 3 of 3 daily checks.
30 / 30
Speed
Typical response 191 ms.
16 / 20
Tool docs
21 of 21 tools have a full description; 42 of 42 inputs have a description.
25 / 25
Stability
16 tool rewrites in 30 days (description or inputs changed, or removed).
0 / 15
Listing
Missing an icon.
9 / 10

How to raise the grade

  • Answer initialize faster: under 150 ms earns full Speed points (yours: 191 ms).
  • Change tool descriptions and inputs rarely, and with a new version: agents trust stable tools.
  • Add an icon to your server.json in the official MCP registry.
Fixed something? Run the same check now. The grade updates at once.

Connect Remote server

Terminal
claude mcp add --transport http api https://api.gaents.com/mcp
~/.cursor/mcp.json
{
  "mcpServers": {
    "api": {
      "url": "https://api.gaents.com/mcp"
    }
  }
}
.vscode/mcp.json
{
  "servers": {
    "api": {
      "type": "http",
      "url": "https://api.gaents.com/mcp"
    }
  }
}
Claude Desktop
Settings > Connectors > Add custom connector
URL: https://api.gaents.com/mcp

Tools 21 found in our last check

  • chat

    Generates a chat completion from an OpenAI-style messages list with the model you name: a gaents/<family> alias (routed across providers, with failover) or an id from GET https://api.gaents.com/v1/models. Use it for open-ended generation, reasoning, summarising, classification o…

  • catalogue

    Lists every module this server sells (chat model router, paid jobs, seller listings) with its path, MCP tool name, price, unit, params and whether it is available now. Free; no input. Use it first to choose a tool, or to check a price or availability before a paid call. Use chat…

  • models_list

    Lists the chat models this server serves: each gaents/<family> alias with its price per prompt and completion token, context length and the curation verdict from our own task tests (best for, superseded on, untested). Free; no input. Use it to pick a model and check its price be…

  • recommend

    Recommends what to call for a task from our own measurements: the rows on the measured frontier (top score or cheapest cost per correct output), each with score, cost per correct, n, date, source and the gaents alias or job to call. Measured tasks now: coding, ner, ner-conll2003…

  • job_transcribe

    Transcribes the speech in one public audio or video file (direct https link to mp3, wav, m4a, ogg, flac, mp4, mov, webm and other FFmpeg formats) to plain text with Whisper small; the language is detected. Use it for podcasts, interviews, calls, lectures or voice notes hosted at…

  • job_transcribe_premium

    Transcribes the speech in one public audio or video file (direct https link to mp3, wav, m4a, ogg, flac, mp4, mov, webm and other FFmpeg formats) to plain text with the larger Whisper large-v3-turbo model; the language is detected. Use it when transcript accuracy matters: names,…

  • job_site_render

    Renders a website in a real headless browser (JavaScript run) and returns the visible text of its start page plus linked contact, about and team pages. Use it to read a company or product site whose content needs JavaScript, or to collect its contact and about pages in one call.…

  • job_youtube_transcript

    Fetches the caption transcript of one YouTube video (uploaded or auto-generated captions), as plain text with the caption language. Use it to read, quote or summarise what is said in a YouTube video. Use job_transcribe instead for a video with no captions (download the media to…

  • job_google_trends

    Fetches Google Trends interest over the last 12 months, worldwide, for one search keyword: weekly relative interest (0-100, 100 = the peak) plus average, peak and latest values. Use it to see whether interest in a topic, product or brand is rising, falling or seasonal. Use job_y…

  • job_youtube_channel

    Lists the latest videos of one YouTube channel, newest first, with title, URL, views, duration and publish time. Use it to monitor a channel's uploads or to pick its videos for transcripts or comments. Use job_youtube_search instead to find videos by topic across all channels. U…

  • job_youtube_comments

    Fetches the top-level comments on one YouTube video (no replies), ranked as YouTube ranks them, with author, likes and reply count. Use it for audience reaction, questions or sentiment on a video. Use job_youtube_transcript instead for what is said in the video itself. Cost: $0.…

  • job_youtube_search

    Searches YouTube for one query and returns the ranked result videos with title, URL, channel, views, duration and date. Use it to find videos on a topic, or to see which channels rank for a query. Use job_youtube_channel instead for one known channel's uploads. Use job_google_tr…

  • job_ocr_tesseract

    Extracts the printed text from one image (PNG, JPEG, TIFF or WebP, sent as base64) with Tesseract OCR 5, English, and reports Tesseract's mean word confidence. Use it for scans, screenshots, photos of printed pages or receipts, and scanned PDF pages saved as images. Use job_doc_…

  • job_doc_to_markdown

    Converts one document (PDF with a text layer, DOCX, PPTX, XLSX, XLS, HTML, CSV, JSON, XML, EPUB, TXT; sent as base64) to Markdown with Microsoft MarkItDown, keeping headings, lists and tables. Use it to feed a document's content to a model or search index as clean Markdown. Use…

  • job_tts_piper

    Speaks English text aloud with Piper TTS and returns a WAV file (base64, 22.05 kHz mono). Voices: en_US-ljspeech-medium (one female voice, public domain) and en_US-libritts_r-medium (904 speakers, CC BY 4.0). Use it for voice-overs, audio versions of short texts and spoken promp…

  • shelf_search

    Searches the results shelf: results we already produced (datasets, transcripts, scored trials, research) sold with their full recipe, measured original cost and hashes. Matches every word of q against titles, summaries, tags and recipe fields; no q lists every item. Use it first…

  • shelf_buy

    Buys one shelf item's OUTPUT (the result files) or its RECIPE (the full recipe: model ids, harness code, settings, prompt, inputs) and returns every file in full with its sha256, a receipt and a signed proof. The price is a fraction of the measured original cost; items with an U…

  • account_balance

    Shows your gaents account: balance in USD, money received by kind (starter credit, paid top-ups, sales), active API keys, seller earnings if you sell, and how to top up. Use it to check you can afford a paid call, or to see your balance after one. Use job_status instead for one…

  • job_status

    Gets the outcome of one past paid call (job, chat, listing run, shelf buy or recommendation) by its proof id: status, amount charged, the ledger legs on your account, usage, and the signed public proof with its check. Jobs here run synchronously; every paid answer carries its pr…

  • seller_statement

    Gets your statement: sales, payable and paid-out earnings, payout key and platform fee, plus every ledger row on your account in the period with its quote and call, and the opening and closing balance. Use it to reconcile sales and earnings, or to audit what your account was cha…

  • listing_create

    Creates a seller listing: your HTTPS endpoint sold per call at the price you set, callable by anyone as tool run_<slug> or POST /v1/run/<slug>; each sale is credited to you minus the platform fee. Posting your own slug again appends a new version. Use it to sell your own API or…

Tool changes seen by our daily check

  • 21 Nov 2026

    Added: account_balance, job_status, listing_create, models_list, recommend, seller_statement. Removed: job_ner_gliner. Description changed: catalogue, chat, job_doc_to_markdown, job_google_trends, job_ocr_tesseract, job_site_render, job_transcribe, job_transcribe_premium, job_tts_piper, job_youtube_channel, job_youtube_comments, job_youtube_search, job_youtube_transcript, shelf_buy, shelf_search.

  • 20 Nov 2026

    Added: shelf_buy, shelf_search.

Renames and new inputs are usually routine. A changed description changes what your agent is told the tool does, and can carry new instructions: review it before you rely on the tool again.

Details

Endpointhttps://api.gaents.com/mcp (Streamable HTTP)
First published19 Nov 2026
Last updated19 Nov 2026
Registry namecom.gaents/api

More ai and media servers See all

Listing data from the official MCP registry. Status and tools from NestDaddy's daily check. Browse all MCP servers.