Step-by-step setup for Claude Desktop, ChatGPT, Cursor, VS Code, the Anthropic API, and any LLM framework. Get live crypto data flowing in minutes.
Native MCP support - 12 crypto tools available directly in Claude.
Sign up free - no credit card needed. Your key is ready instantly.
Download Claude Desktop if you don't have it yet, then install and open it.
Go to Claude Desktop → Settings → Developer → Edit Config.
This opens claude_desktop_config.json.
Add the "guavy-crypto"
entry inside the "mcpServers"
object. If you already have other servers, keep them — just add this alongside.
Replace YOUR_GUAVY_API_KEY with your key.
{
"mcpServers": {
"guavy-crypto": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://data.guavy.com/mcp?key=YOUR_GUAVY_API_KEY"]
}
}
}
Requires Node.js installed
(Mac: brew install node
· Windows: winget install OpenJS.NodeJS
· Linux: sudo apt install nodejs npm).
After restarting, you will see a hammer icon in the chat input bar. Click it — you should see 12 Guavy tools listed.
Ask Claude: "What is the current BTC trade signal?"
Claude will call get_current_action live
and return a real Buy/Sell/Hold response.
Import Guavy into any Custom GPT via OpenAPI Actions - all endpoints in 60 seconds
Sign up free and copy your API key from the dashboard.
Go to chatgpt.com/gpts → Create a GPT → Configure → Actions → Create new action.
In the schema field, enter this URL - GPT will auto-import all Guavy endpoints:
https://data.guavy.com/openapi.json
Under Authentication, set Type to API Key, Auth type to Bearer, and paste your Guavy API key.
Paste into the Instructions field for best results:
Use the Guavy Action for every crypto data request. Do not use web search or answer from memory for prices, signals, sentiment, news, or recommendations. Always call the Guavy API for live data - never guess. When asked about a coin, check sentiment history and current action before responding. If a Guavy action fails, say that live data was unavailable rather than guessing.
Under Capabilities, turn off Web Search so your GPT always uses live Guavy data instead of searching the web.
Ask your GPT: "What are the current crypto buy recommendations?"
Add Guavy as an MCP server in Cursor or VS Code GitHub Copilot
Go to Cursor Settings → MCP → Add MCP Server.
Guavy CryptoSSEhttps://data.guavy.com/mcp
Authorization: Bearer YOUR_GUAVY_API_KEY
In Cursor Chat: "@guavy-crypto What is the sentiment trend for ETH this week?"
Add this file (replace YOUR_GUAVY_API_KEY):
{
"servers": {
"guavy-crypto": {
"type": "sse",
"url": "https://data.guavy.com/mcp",
"headers": {
"Authorization": "Bearer YOUR_GUAVY_API_KEY"
}
}
}
}
Copilot will auto-discover the 12 Guavy tools. Use them in the Copilot Chat sidebar.
Define Guavy as native tools in your Claude API calls - no wrappers needed
pip install anthropic requests
import anthropic import requests client = anthropic.Anthropic(api_key="YOUR_ANTHROPIC_KEY") GUAVY_KEY = "YOUR_GUAVY_API_KEY" GUAVY_BASE = "https://data.guavy.com/api/v1" tools = [ { "name": "get_current_action", "description": "Get the current trade signal (Buy/Sell/Hold/Stay) for a coin", "input_schema": { "type": "object", "properties": { "symbol": {"type": "string"}, "strategy": {"type": "string", "enum": ["aggressive", "conservative"]} }, "required": ["symbol", "strategy"] } }, { "name": "get_market_summary", "description": "Get today's AI-generated crypto market summary", "input_schema": {"type": "object", "properties": {}} } ] def call_guavy(tool_name, tool_input): headers = {"Authorization": f"Bearer {GUAVY_KEY}"} if tool_name == "get_current_action": r = requests.get( f"{GUAVY_BASE}/trades/get-current-action/{tool_input['symbol']}/{tool_input['strategy']}", headers=headers ) elif tool_name == "get_market_summary": r = requests.get(f"{GUAVY_BASE}/newsroom/get-market-summary", headers=headers) return r.json() messages = [{"role": "user", "content": "What is the BTC signal and overall market today?"}] response = client.messages.create( model="claude-opus-4-5", max_tokens=1024, tools=tools, messages=messages ) while response.stop_reason == "tool_use": tool_results = [] for block in response.content: if block.type == "tool_use": result = call_guavy(block.name, block.input) tool_results.append({ "type": "tool_result", "tool_use_id": block.id, "content": str(result) }) messages += [ {"role": "assistant", "content": response.content}, {"role": "user", "content": tool_results} ] response = client.messages.create( model="claude-opus-4-5", max_tokens=1024, tools=tools, messages=messages ) print(response.content[0].text)
Two patterns: llms.txt for spec injection, or native MCP via LangChain adapter
pip install anthropic requests
Fetches the API spec at runtime — always current, no hardcoding. Works with any LLM or framework that accepts a system prompt.
import os, requests, anthropic GUAVY_KEY = os.environ["GUAVY_API_KEY"] ANTHROPIC_KEY = os.environ["ANTHROPIC_API_KEY"] # Pull the live spec — updates automatically when Guavy adds endpoints spec = requests.get("https://data.guavy.com/llms.txt").text system = f"""You are a crypto analyst. You have access to live market data \ via the Guavy API. Use it to answer every question — never guess prices or signals. {spec} Auth: Authorization: Bearer {GUAVY_KEY} """ client = anthropic.Anthropic(api_key=ANTHROPIC_KEY) response = client.messages.create( model="claude-opus-4-5", max_tokens=1024, system=system, messages=[{"role": "user", "content": "What's the BTC trade signal right now?"}] ) print(response.content[0].text)
pip install langchain-mcp-adapters langchain-anthropic langgraph
The agent auto-discovers all 12 Guavy tools and decides which to call based on your question.
import os, asyncio from langchain_mcp_adapters.client import MultiServerMCPClient from langchain_anthropic import ChatAnthropic from langgraph.prebuilt import create_react_agent GUAVY_KEY = os.environ["GUAVY_API_KEY"] async def ask(agent, question): result = await agent.ainvoke({"messages": question}) print(f"\n> {question}") print(result["messages"][-1].content) async def main(): async with MultiServerMCPClient({ "guavy": { "url": "https://data.guavy.com/mcp", "transport": "sse", "headers": {"Authorization": f"Bearer {GUAVY_KEY}"} } }) as client: tools = await client.get_tools() agent = create_react_agent(ChatAnthropic(model="claude-opus-4-5"), tools) # Each call picks the right tool automatically await ask(agent, "What's the overall market doing today?") await ask(agent, "Which coins have a buy signal right now?") await ask(agent, "What's the ETH sentiment trend over the past 2 weeks?") asyncio.run(main())
// Ready
Free sandbox forever. No credit card required. All 350+ coins available via MCP.
Disclaimer: Guavy is a data and market intelligence provider, not an investment advisor. The information, signals, and market analysis provided by the Guavy mobile application, API, and related services are for informational purposes only and are not intended as financial advice, investment recommendations, or an endorsement of any particular trading strategy. Cryptocurrency trading is highly volatile, carries significant risk, and may not be suitable for all investors. Past performance is not indicative of future results. Users should consult with a qualified financial professional before making any investment decisions. Guavy makes no guarantee of trading profits or financial returns.
Real-Time Market Intelligence for Apps, Funds & Agents
Location
729 55 Ave SW
Calgary AB T2V 0G4
Canada
Get Your Daily Crypto Market Update
Receive the most important news, market sentiment, and prices for leading instruments. Delivered straight to your inbox.
© 2025 Guavy Inc