A2A INTEGRATION · CUSTOM TOOL

Use GitDealFlow A2A with AutoGen

Microsoft's multi-agent conversation framework. Researchers and enterprise builders running multi-agent debates and tool-use loops.

Crunchbase API: $20K/yr. GitDealFlow A2A: free, no signup.

AutoGen's FunctionTool wraps any Python callable. Wrap our A2A endpoint as a FunctionTool and any AssistantAgent in a GroupChat can pull live signals as part of a multi-agent debate ('the scout argues for inclusion, the contrarian argues against, the methodology agent cites SSRN').

Install

pip install autogen-agentchat autogen-ext requests

FunctionTool wrapper

from autogen_agentchat.agents import AssistantAgent
from autogen_ext.tools import FunctionTool
from autogen_ext.models.openai import OpenAIChatCompletionClient
import requests

A2A_URL = "https://signals.gitdealflow.com/api/a2a"

def gitdealflow_query(skill: str, args: dict | None = None) -> dict:
    """Call GitDealFlow A2A for VC engineering signals.

    skill: get_trending_startups | search_startups_by_sector |
           get_startup_signal | get_signals_summary | get_methodology
    """
    body = {
        "jsonrpc": "2.0", "id": 1,
        "method": "message/send",
        "params": {"message": {"role": "user", "parts": [
            {"kind": "data", "data": {"skill": skill, "args": args or {}}},
        ]}},
    }
    return requests.post(A2A_URL, json=body, timeout=15).json()

scout_tool = FunctionTool(gitdealflow_query, description="VC engineering signals.")

scout = AssistantAgent(
    name="scout",
    model_client=OpenAIChatCompletionClient(model="gpt-4o-mini"),
    tools=[scout_tool],
    system_message="You surface live engineering signals via gitdealflow_query.",
)

What you can ask

GroupChat: scout + contrarian + memo-writer debate the top breakout.
Two-agent loop: planner asks scout for fintech trending, scout returns, planner picks top 3.
Critic agent challenges scout's pick by asking for methodology citation.
Reflection loop: scout proposes, critic checks, scout revises.

Gotchas

AutoGen v0.4+ uses async tools — wrap requests with asyncio.to_thread or use aiohttp instead.
GroupChat token cost compounds quickly; cap rounds with max_turns and have the scout summarize the A2A payload before passing it onward.

References

Try it without setup

The interactive playground lets you send live JSON-RPC requests against the A2A endpoint with no install, no auth. Pick a skill, hit send, see the response.

Full launch story: I made my VC deal flow callable by Claude.

Other framework integrations

FunctionTool wrapper

from autogen_agentchat.agents import AssistantAgent from autogen_ext.tools import FunctionTool from autogen_ext.models.openai import OpenAIChatCompletionClient import requests A2A_URL = "https://signals.gitdealflow.com/api/a2a" def gitdealflow_query(skill: str, args: dict | None = None) -> dict: """Call GitDealFlow A2A for VC engineering signals. skill: get_trending_startups | search_startups_by_sector | get_startup_signal | get_signals_summary | get_methodology """ body = { "jsonrpc": "2.0", "id": 1, "method": "message/send", "params": {"message": {"role": "user", "parts": [ {"kind": "data", "data": {"skill": skill, "args": args or {}}}, ]}}, } return requests.post(A2A_URL, json=body, timeout=15).json() scout_tool = FunctionTool(gitdealflow_query, description="VC engineering signals.") scout = AssistantAgent( name="scout", model_client=OpenAIChatCompletionClient(model="gpt-4o-mini"), tools=[scout_tool], system_message="You surface live engineering signals via gitdealflow_query.", )