FetchPan

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Download links from file hosts and Chinese cloud drives (Rapidgator, Baidu, Quark) for AI agents.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "FetchPan": {
      "url": "https://fetchpan.com/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (9)

• list_supported_sites read-only

  List the sites FetchPan can download from, with live status (working, slow, down). `main_sites` are the big ones (Baidu, Quark, Rapidgator ...). `file_hosts` are the other file hosts. `free` tells whether a free download is possible there or only a paid one.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• check_link read-only

  Is this link from a site FetchPan supports, and is that site working right now? Does not use any quota. Call it before `download` when you are not sure.

  Parameters:

  • url (required) — string — The link the user wants to download

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• start_download

  Start fetching a link. Returns a job id; poll it with `get_job`. Uses no free file yet.

  Parameters:

  • url (required) — string — Share or file link from a supported site
  • password — Extraction code / password, if the share has one

  Hints: read-only no · destructive no · idempotent no · open-world yes

• get_job read-only

  Status of a job: queued, running, succeeded or failed. When it has succeeded it also says whether a free download is possible (`free`) and what a fast one would cost in traffic bytes (`fast`).

  Parameters:

  • job_id (required) — string — Job id from start_download or download

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• get_download_link

  Get the download URL of a succeeded job. A free link uses one of the account's free files for the day; a fast link uses traffic from the account balance. Asking again for the same job and tier returns the same link.

  Parameters:

  • job_id (required) — string — A succeeded job
  • tier — string — "free" (limited per day and speed) or "fast" (uses traffic balance)

  Hints: read-only no · destructive no · idempotent yes · open-world yes

• download

  Fetch a link and return a download URL, in one call. Waits up to `wait_seconds` for the file to be ready. If it is not ready yet, the result has status "running" and a `job_id`: call `get_job`, then `get_download_link`.

  Parameters:

  • url (required) — string — Share or file link from a supported site
  • tier — string — "free" or "fast"
  • password — Extraction code / password, if any
  • wait_seconds — integer — How long to wait for the file to be ready

  Hints: read-only no · destructive no · idempotent no · open-world yes

• get_account read-only

  The account behind the token: email, traffic balance in bytes, free files per day, and the last downloads.

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• list_packages read-only

  Traffic packs that can be bought (code, size in bytes, price in USDT).

  Hints: read-only yes · destructive not declared · idempotent not declared · open-world no

• create_payment_link

  Create a payment page for a traffic pack and return its URL. Give the URL to the user: only a person can pay it. Nothing is charged by calling this.

  Parameters:

  • package_code (required) — string — A code from list_packages

  Hints: read-only no · destructive no · idempotent no · open-world yes

Publisher website

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.