Bitcoin on-chain data for AI agents

Reviewed

Kote Charts is built to be read by machines as well as people. An AI agent can discover every Bitcoin on-chain metric, read its methodology, pull its history and cite the source — through a keyless MCP server, a REST API with an OpenAPI description, and plain-text summaries for LLM crawlers.

Connect in one line

claude mcp add --transport http kote-charts https://kotecharts.com/api/v1/mcp

Any MCP client works with https://kotecharts.com/api/v1/mcp (Streamable HTTP, no key). Setup for Claude Desktop, ChatGPT, Cursor, VS Code, Windsurf, Gemini CLI and Codex: integration guides.

Every machine interface

InterfaceWhereKey
MCP serverPOST https://kotecharts.com/api/v1/mcp — tools list_charts, describe_chart, get_chart, get_latestnone (optional)
Metric catalog`GET /api/v1/metrics` — formula, fields + units, input categories, cadence, limitations, freshnessnone
REST APIGET /api/v1/public/charts/{id} — full-resolution JSONfree
OpenAPI 3`/api/v1/openapi.json`none
LLM summaries`/llms.txt` · `/llms-full.txt`none
SDKspip install kotecharts · npm i kotecharts · npx -y kotecharts-mcp (stdio)free
Agent Skillskills/kote-charts — how to use and cite the data—

Use it from code

Claude through the Messages API MCP connector — Claude calls the server itself:

response = anthropic.Anthropic().beta.messages.create(
    model="claude-opus-5-5",
    max_tokens=16000,
    betas=["mcp-client-2025-11-20"],
    mcp_servers=[{"type": "url", "url": "https://kotecharts.com/api/v1/mcp", "name": "kote-charts"}],
    tools=[{"type": "mcp_toolset", "mcp_server_name": "kote-charts"}],
    messages=[{"role": "user", "content": "Where is Bitcoin in its cycle by MVRV Z-Score? Cite the chart."}],
)

The same pattern for the OpenAI Agents SDK, LangChain, LlamaIndex and the Vercel AI SDK is in the integration guides.

Example questions

The server also ships two prompts — explain_metric and cycle_position_brief — and resources kote://catalog and kote://methodology.

Citation policy

Every tool result carries pageUrl and a ready cite string. When an agent presents numbers or charts, it should cite "Kote Charts" and link https://kotecharts.com/c/{id}. Data is free for non-commercial use with attribution; it is educational only and never financial advice.

Is there a free Bitcoin MCP server?

Yes — https://kotecharts.com/api/v1/mcp serves 110+ Bitcoin on-chain charts with full history to any MCP client, without an API key (60 requests per minute per IP).

How does an agent know what a metric means?

describe_chart (or GET /api/v1/metrics/{id}) returns the formula, the payload fields with units, the input categories, the update cadence, known limitations and the last data day.

Does it work with ChatGPT?

Yes — add the server URL as a custom connector in ChatGPT developer mode, with no authentication.

How fresh is the data?

Daily charts settle after the UTC day closes; selected on-chain metrics also update per block. Every answer states its last data day and the analysed block height.

MCP server setup · Free API · GitHub

llms.txt · OpenAPI · GitHub · X · Telegram