CryptoQuant wants you to stop squinting at dashboards and start asking questions instead.
The on-chain analytics firm has launched a Model Context Protocol (MCP) server that connects its data to large language models. In practice, a trader can ask an AI assistant about Bitcoin exchange flows the same way they’d ask it to summarize an email.
What CryptoQuant actually built
MCP is a protocol that lets LLMs and AI agents reach outside their training data and pull from external sources.
CryptoQuant’s server exposes over 245 metrics covering real-time on-chain, market, and derivatives data. Those include MVRV, SOPR, exchange flows, and whale activity.
For the uninitiated, MVRV and SOPR are ratios analysts use to gauge whether holders are sitting on gains or losses. Exchange flows track coins moving on and off trading platforms. Whale activity watches what the largest wallets are doing.
The server handles natural-language queries in both English and Korean. That second language is a notable choice given how active Korean traders are in crypto markets.
Asset coverage is broad. The server supports Bitcoin (BTC), Ethereum (ETH), TRX, XRP, various stablecoins, and ERC-20 tokens.
CryptoQuant runs a remote server endpoint at mcp.cryptoquant.com. Users can authenticate through OAuth 2.0 with PKCE or with an API key.
Compatible clients include Claude.ai and Cursor.
The business model behind the bot
This isn’t a token launch, and there’s no new coin attached. The MCP server is an extension of CryptoQuant’s existing paid API access.
Pricing follows the company’s established credit-based structure. That means AI queries draw from the same metered system as conventional API usage.
CryptoQuant’s free Basic tier lets users start experimenting with the AI functionality before paying anything.
The server first appeared in beta earlier in 2026 and remains actively maintained. CryptoQuant has highlighted beta features and related research content from March through October 2026.
Where CryptoQuant fits
CryptoQuant has operated as an on-chain data provider since at least 2019, serving traders and institutions.
What this means for traders and analysts
The most immediate impact is on accessibility. Someone who understands markets but can’t write Python now has a path to the same metrics a quant desk uses.
For professional users, the appeal is speed. Asking an agent to compare SOPR trends across assets is faster than building the comparison by hand, assuming the AI interprets the request correctly.
That assumption deserves attention. An LLM fetching accurate data is one thing. An LLM correctly framing what that data means is another. Users will still need enough context to spot when a confident answer is confidently wrong.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.