Kick MCP server

Short answer

ScrapeIO runs a hosted MCP server that gives Claude, Cursor, Codex and other AI agents read-only access to 2 Kick endpoints through its call_endpoint tool, for example /v1/kick/clip. Add https://scrapeio.com/mcp with your API key; there is nothing to install or run. Calls cost the same credits as the API and new accounts get 100 free credits.

Server URL	`https://scrapeio.com/mcp`
Transport	Streamable HTTP, hosted. Nothing to install.
Auth	`x-api-key` header or Bearer token
Kick tools	2 (0 listed directly, 2 through call_endpoint)
Access	Read-only public data
Price	Same credits as the API. 100 free on signup.

Tools

Tool	What it returns	Credits
`call_endpoint` `/v1/kick/clip`	Clip	1
`call_endpoint` `/v1/kick/clip/transcript`	Clip Transcript	1

Set it up

Get a key at scrapeio.com/login and put it in your shell profile as export SCRAPEIO_API_KEY=sk_.... Then add the server to your agent:

Claude Code

Run once in a terminal. Or install the plugin, which adds the skill too: /plugin marketplace add xhoantran/scrapeio-agent, then /plugin install scrapeio@scrapeio.

claude mcp add --transport http scrapeio https://scrapeio.com/mcp \
  --header "x-api-key: $SCRAPEIO_API_KEY"

Cursor

Add to ~/.cursor/mcp.json (or .cursor/mcp.json in a project).

{
  "mcpServers": {
    "scrapeio": {
      "url": "https://scrapeio.com/mcp",
      "headers": { "x-api-key": "${env:SCRAPEIO_API_KEY}" }
    }
  }
}

Codex

Add to ~/.codex/config.toml. Codex sends the key as a Bearer token, which the server accepts.

[mcp_servers.scrapeio]
url = "https://scrapeio.com/mcp"
bearer_token_env_var = "SCRAPEIO_API_KEY"

Gemini CLI, VS Code, Windsurf, Hermes and other clients are on the agents page. All 33 platforms come with the same server, so one setup covers Kick and the rest.

Prompts to try

Get the transcript of this Kick video and give me the five main points with quotes.

The agent picks the tools, pages through results when it needs more, and reports the numbers it got back. Each answer uses the same field names on every platform, so a prompt that mixes Kick with TikTok or YouTube works the same way.

Why a hosted MCP server

No scraping inside the agent. Browsing Kick from an agent runs into login walls and changing pages. A tool call returns clean JSON in seconds.
One schema. handle, followers, plays, likes and created_at mean the same thing on every platform, so the model does not have to translate.
Pay per call. No monthly plan for an agent that runs now and then. Failed and cached calls are free.

Questions

Is the Kick MCP server free?

Connecting is free and new accounts get 100 free credits. After that you buy credit packs ($47 for 25,000 or $497 for 500,000). Credits never expire, and failed or cached calls cost nothing.

Do I need a Kick account or developer key?

No. The server returns public Kick data with your ScrapeIO key only. There is no Kick app to register and no OAuth.

Can the agent post, like or send messages?

No. Every tool is read-only and marked readOnlyHint, so clients can run them without a write warning.

Which clients work?

Any client that speaks Streamable HTTP: Claude Code, Cursor, Codex, Gemini CLI, VS Code agent mode, Windsurf, Hermes and others. It is also in the official MCP registry as com.scrapeio/scrapeio.

Why are some Kick tools reached through call_endpoint?

To keep the tool list short in every client's context, the core tools are listed directly and the rest are one call away: the agent finds them with list_endpoints and runs them with call_endpoint. It pays the same credits either way.

Prefer code? The same data is on the Kick API page.

100 free credits, no card. Credits never expire.

Get API key