Integration guide

apifare for CrewAI

Give CrewAI one prepaid key for every API it needs. There is no separate provider account or provider API key for eligible capabilities, and it all runs under hard caps and a kill switch.

Get a key

1 · Get a key

Create an account at /start: email in, token out (shown once). Save the token somewhere safe.

2 · Add apifare to CrewAI

Give your agent apifare through its mcps field with MCPServerHTTP. The token is read from the APIPAY_TOKEN environment variable and sent as an Authorization header. Keep the token out of the URL.

pip install crewai

import os
from crewai import Agent, Crew, Task
from crewai.mcp import MCPServerHTTP

agent = Agent(
    role="Research Analyst",
    goal="Answer questions using live APIs",
    backstory="Analyst with pay-per-call API access via apifare",
    mcps=[
        MCPServerHTTP(
            url="https://apifare.com/mcp",
            headers={"Authorization": f"Bearer {os.environ['APIPAY_TOKEN']}"},
            streamable=True,
        )
    ],
)
task = Task(description="What's the weather in London right now?",
            expected_output="A one-paragraph answer.", agent=agent)
print(Crew(agents=[agent], tasks=[task]).kickoff())

Set APIPAY_TOKEN and your model provider's key, then run it. For a local server instead, MCPServerStdio(command="npx", args=["-y", "apifare-mcp"], env={...}) takes the same token.

3 · Try it

Ask your crew to search the web for X, then to fetch a URL, then what's the weather in London. Every call is debited from your balance and lands on your signed, exportable audit chain.

4 · Stay in control

Set daily or monthly caps per agent, and keep the kill switch ready. Both are checked before a call is dispatched. When the balance runs out, your agent gets an HTTP 402 with a top-up link it can hand you; the error catalog lists every stop and what unblocks it.

What CrewAI can call

Everything on the key: search, filings, registries, weather, and ready-to-run tools, plus any provider key you vault yourself. Exact per-call prices are on /capabilities.

Config checked against CrewAI's MCP docs on 2026-09-13. Markdown version: /integrations/crewai.md.

More integrations

Claude Code · Claude Desktop · Cursor · Windsurf · Zed · Cline · GitHub Copilot (VS Code) · OpenAI Codex CLI · Gemini CLI · Hermes Agent · Kimi CLI · Qwen Code · JetBrains (Junie) · Warp · Raycast · n8n · Dify · LangChain · LangGraph · Vercel AI SDK · LobeHub · AnythingLLM · OpenClaw · OpenAI Agents SDK · LlamaIndex · Mastra · Goose · LibreChat

Model guides

Calling a model's API directly instead? Give it apifare tools with a plain tool loop: DeepSeek · Kimi · Qwen · OpenAI GPT · Gemini.