mcpField
C
crypto-pulse-mcp

crypto-pulse-mcp

Local
github.com/markankamaransheogarat-sysFinance0Updated 1w ago

About crypto-pulse-mcp

crypto-pulse-mcp is a read-only MCP server for AI assistants that returns structured crypto market data from public APIs without API keys or trading access. It covers exchange prices and derivatives alongside market rankings, DeFi TVL, and stablecoin supply.

Key features

  1. 01Spot prices and exchange gaps
  2. 02Funding, basis, and open interest
  3. 03Market rankings and breadth
  4. 04DeFi chain and protocol TVL
  5. 05Stablecoin supply changes

Use cases

  1. 01Compare a coin’s Bybit and Binance prices
  2. 02Check funding and open-interest conditions
  3. 03Review chain and protocol TVL
  4. 04Track stablecoin supply trends

MCP support

Exposes read-only crypto data tools to MCP clients, including price, derivatives, market overview, DeFi TVL, and stablecoin queries.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"markankamaransheogarat-sys-crypto-pulse": {
"command": "npx",
"args": [
"-y",
"crypto-pulse-mcp"
]
}
}
}
Run commandnpx -y crypto-pulse-mcp

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmcrypto-pulse-mcp0.1.0

Tools

5

Saved tool list · updated 1w ago

ParameterTypeRequiredDescription
symbolstringYesCoin or pair, e.g. BTC, ETH/USDT, SOLUSDT

Details

From the project README

crypto-pulse-mcp

Read-only crypto market data for AI assistants, over the Model Context Protocol. Ask Claude (or any MCP client) about prices, funding, open interest, DeFi TVL and stablecoin flows and get structured answers from public APIs. No API keys. No trading. No account access.

ToolWhat it answers
get_priceSpot price, 24h change/range/volume on Bybit and Binance, and the gap between the two
get_derivativesPerpetual funding rate (also annualised), basis, open interest and its ratio to volume
get_market_overviewTop coins by market cap or volume, 24h change, distance from ATH, market breadth (CoinGecko)
get_defi_tvlTop chains by TVL, one chain with its top protocols (e.g. Solana, Stellar), or one protocol (DefiLlama)
get_stablecoinsLargest USD stablecoins, supply share and 1d/7d/30d supply change, a rough proxy for money entering or leaving crypto

Example prompts once it is connected:

  • "Is SOL trading at a premium on Bybit compared with Binance right now?"
  • "What is BTC funding doing? Are longs crowded?"
  • "How big is Stellar DeFi, and which protocols lead it?"
  • "Is stablecoin supply growing or shrinking this week?"

Example

Asked in Claude Code with the server connected (snapshot, 22 Oct 2026):

Is SOL at a premium on Bybit vs Binance? What is BTC funding doing? How big is Stellar DeFi?

  • SOL (get_price): no premium. $103.82 on Bybit, $103.83 on Binance, about 0.01% apart.
  • BTC funding (get_derivatives): slightly negative, -0.0041% per 8h on Bybit (about -4.5% a year), so shorts are paying longs. Longs do not look crowded.
  • Stellar DeFi (get_defi_tvl): about $241M TVL, #18 among chains. Top protocol: Blend Pools V2, about $140M.

Install

Requires Node.js 20 or newer. The quickest way is npx, no clone needed:

claude mcp add crypto-pulse -- npx -y crypto-pulse-mcp

On Windows, run it from cmd (PowerShell swallows the --) and wrap npx so Claude Code can start it:

claude mcp add crypto-pulse -- cmd /c npx -y crypto-pulse-mcp

Add -s user to make it available in every folder, not just the current project.

Published on npm and in the official MCP Registry as io.github.markankamaransheogarat-sys/crypto-pulse-mcp.

To run it from source instead:

git clone https://github.com/markankamaransheogarat-sys/crypto-pulse-mcp.git
cd crypto-pulse-mcp
npm install
npm run build
Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "crypto-pulse": {
      "command": "npx",
      "args": ["-y", "crypto-pulse-mcp"]
    }
  }
}

On Windows use "command": "cmd" with "args": ["/c", "npx", "-y", "crypto-pulse-mcp"].

Claude Code
claude mcp add crypto-pulse -- node /absolute/path/to/crypto-pulse-mcp/dist/index.js

Any other MCP client that can launch a stdio server works the same way.

Details worth knowing

  • Symbols. BTC, ETH/USDT, SOLUSDT and doge-usdt all work; a bare coin means a USDT pair.
  • One exchange down is not a failure. get_price and get_derivatives query both exchanges in parallel. If one is unreachable (Binance is geo-blocked in some countries), you still get the other, with an errors field saying so. An unknown symbol is reported as "not listed", not as an outage.
  • Caching. Responses are cached briefly (5 s for exchange tickers, 60 s for CoinGecko, 5 min for DefiLlama) to stay inside free-tier rate limits. Simultaneous identical requests share one call.
  • Funding sign. Positive funding means longs pay shorts. annualizedFundingPct assumes today's rate holds for a year, which it will not; read it as a scale, not a forecast.
  • DefiLlama CEX entries (exchange wallet balances) are excluded from per-chain protocol rankings.
  • Not financial advice. These are snapshots of public data and can be seconds to minutes old.

Development

npm run typecheck   # tsc --noEmit over src and tests
npm test            # vitest, all HTTP mocked
npm run build       # compile to dist/
node scripts/smoke.mjs   # live check: starts the built server and calls every tool against the real APIs

Layout: src/sources/* talk to one API each and return normalised shapes, src/tools/* hold the logic and register one MCP tool each, src/http.ts is the shared client (timeout, bounded retries, TTL cache), src/symbols.ts normalises symbols.

Data sources

Bybit v5 · Binance spot and USDⓈ-M futures · CoinGecko · DefiLlama. This project is not affiliated with any of them; respect their terms and rate limits.

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.