MCP

MeowLoad ships a hosted MCP server (Model Context Protocol, Streamable HTTP). Once it is set up, you can just say "extract the video and images from this link" in Claude Code, Codex or Cursor and the agent will call the MeowLoad API — no code required. It uses the same API key, the same billing and the same rate-limit quota as the REST API.

Endpoint: https://api.meowload.net/mcp

Authentication: Authorization: Bearer <your API key>, the same key as the REST API.

Before you start

Get your API key from the Developer Console.

The <your API key> in every config below is a placeholder that you must replace with your own real key. Pasting it as-is does not fail at configuration time, but the client will get a 401 as soon as it connects, and no MeowLoad tools will show up.

Claude Code

claude mcp add --transport http meowload https://api.meowload.net/mcp --header "Authorization: Bearer <your API key>"

This only applies to the directory you ran it in. Add -s user to make it available in every project.

Codex

# ~/.codex/config.toml
[mcp_servers.meowload]
url = "https://api.meowload.net/mcp"
bearer_token_env_var = "MEOWLOAD_API_KEY"

bearer_token_env_var takes an environment variable name: on startup Codex reads the value of MEOWLOAD_API_KEY and sends it as Authorization: Bearer <value>. Set the variable in your shell first (for example export MEOWLOAD_API_KEY=<your API key>), and the key never lands in the config file.

If you would rather not use an environment variable, put the header directly in the config:

# ~/.codex/config.toml
[mcp_servers.meowload]
url = "https://api.meowload.net/mcp"
http_headers = { "Authorization" = "Bearer <your API key>" }

~ is your home directory; create the file if it does not exist. Restart Codex afterwards.

Generic JSON config (Cursor and others)

Cursor reads ~/.cursor/mcp.json (global) or .cursor/mcp.json in the project root (that project only). For other clients, check their docs for the MCP config file location.

{
  "mcpServers": {
    "meowload": {
      "url": "https://api.meowload.net/mcp",
      "headers": { "Authorization": "Bearer <your API key>" }
    }
  }
}

Restart the client afterwards.

Verify the connection

In Claude Code, run:

claude mcp list

Seeing meowload listed as connected means you are set. /mcp inside a session works too. Other clients show the server in their MCP settings panel. If the five tools below are listed, it is working.

Available tools

The three extract tools cost the same as the REST endpoints and are charged only on success; get_credits and get_docs are free.

The three extract tools return exactly the same response as the corresponding endpoint in the API reference, and the field descriptions live there.

The extract tools return direct media links and metadata, not the files themselves. To actually save something locally, have your agent download the link it got back; most clients can do that directly. Links are short-lived, so download promptly.

extract_playlist fetches one page per call. When has_more is true in the result, call it again with next_cursor as the cursor argument to get the next page.

get_docs takes a page argument with one of quickstart, migration, errors or mcp, and a lang of zh or en, so the agent can look up field meanings and error codes on its own.

Credits

Your balance, top-ups and alert settings are all in the Developer Console. When you run out, the extract tools return Insufficient credits together with the top-up URL. The billing rules are listed there as well.

Troubleshooting

Next steps

• Quickstart: call the same API over REST
• API reference: full endpoint docs with a try-it console
• Errors: what the error codes returned by the tools mean

Try telling your agent: "Extract the 1080p link and the English subtitles of this YouTube video: https://..."