Onboarding: AutoGen

Add persistent memory to your AutoGen agents. The adapter captures multi-agent conversations and injects past context into reply hooks.

Best path: Gateway (for LLM calls) or Python SDK adapter (for agent-level capture) Time: 2–5 minutes Reliability: ~99% (Gateway) / ~80% (adapter)

---

Option A: Gateway (Recommended for LLM calls)

If your AutoGen agents use OpenAI or another provider for LLM calls, the Gateway is the simplest integration:

from openai import OpenAI

client = OpenAI(
    base_url="https://api.hippocortex.dev/v1",
    api_key="hx_live_...",
    default_headers={"X-LLM-API-Key": "sk-..."},
)

Every LLM call gets memory automatically. ~99% reliability with graceful fallback.

---

Option B: SDK Adapter (For agent-level capture)

Step 1: Install

pip install hippocortex[autogen]

Step 2: Set your API key

export HIPPOCORTEX_API_KEY=hx_live_...

Get a key at dashboard.hippocortex.dev if you do not have one.

Step 3: Wrap your agents

from autogen import AssistantAgent, UserProxyAgent
from hippocortex.adapters import autogen as hx_autogen

# Create your agents normally
assistant = AssistantAgent("assistant", llm_config={...})
user_proxy = UserProxyAgent("user")

# Add memory with one line
assistant = hx_autogen.wrap(assistant, api_key="hx_live_...")

# Use exactly as before
user_proxy.initiate_chat(assistant, message="Deploy to staging")

Step 4: Verify

Check dashboard.hippocortex.dev to see captured conversations between agents.

---

What gets captured

• Inter-agent messages
• Function calls and results
• Conversation turns
• Final outcomes

Note on multi-agent setups

If you have multiple agents, wrap only the primary assistant to avoid duplicate captures. The reply hook captures both sides of each conversation turn.

Auto-compile

Knowledge compilation runs automatically after every 10 captures. You do not need to call learn() manually.