Python Chat Application with Memory

pip install papr_memory fastapi uvicorn

import os
from fastapi import FastAPI
from pydantic import BaseModel
from papr_memory import Papr

app = FastAPI()
client = Papr(x_api_key=os.environ.get("PAPR_MEMORY_API_KEY"))


class ChatRequest(BaseModel):
    session_id: str
    external_user_id: str
    message: str


@app.post("/chat")
def chat(req: ChatRequest):
    client.messages.store(
        sessionId=req.session_id,
        role="user",
        content=req.message,
        external_user_id=req.external_user_id,
        process_messages=True,
    )

    compressed = client.messages.compress_session(session_id=req.session_id)

    memory_results = client.memory.search(
        query=f"Retrieve context relevant to: {req.message}",
        external_user_id=req.external_user_id,
        enable_agentic_graph=True,
    )

    # Replace with your model provider call.
    assistant_reply = (
        "Stub response. Connect your model here using "
        "compressed context and memory search results."
    )

    client.messages.store(
        sessionId=req.session_id,
        role="assistant",
        content=assistant_reply,
        external_user_id=req.external_user_id,
        process_messages=True,
    )

    return {
        "reply": assistant_reply,
        "context_for_llm": compressed.context_for_llm,
        "memory_count": len(memory_results.data.get("memories", []))
        if hasattr(memory_results, "data")
        else None,
    }

uvicorn app:app --reload

curl -X POST http://127.0.0.1:8000/chat \
  -H "Content-Type: application/json" \
  -d '{
    "session_id": "session_001",
    "external_user_id": "user_123",
    "message": "I prefer weekly reports by email."
  }'