Sqlite session persistence

sqlite-session-persistence

Use SQLiteSession(session_id, db_path="...") to persist conversation history to a local SQLite database. History survives process restarts and can be shared across multiple processes reading the same file.

When to apply

Single-server deployments, CLI tools, or development environments where you need persistent multi-turn conversation without a separate database service.

Core snippet

import asyncio
from agents import Agent, Runner, SQLiteSession

agent = Agent(name="Assistant", instructions="Reply very concisely.")

async def main():
    session_id = "user_123_conversation"
    session = SQLiteSession(session_id, db_path="conversations.db")

    # Turn 1 — new conversation
    result = await Runner.run(
        agent,
        "What city is the Golden Gate Bridge in?",
        session=session,
    )
    print(f"Turn 1: {result.final_output}")

    # Turn 2 — history loaded from SQLite
    result = await Runner.run(agent, "What state is it in?", session=session)
    print(f"Turn 2: {result.final_output}")

asyncio.run(main())
# After process restart, run again with the same session_id — history is preserved

Key notes

• Default db_path is ~/.agents_sessions.db if not specified
• Session ID is the lookup key; use a stable per-user or per-conversation ID
• SQLite is single-writer; for multi-process writes, use RedisSession instead
• Install: pip install openai-agents (SQLite included); Redis: pip install 'openai-agents[redis]'
• History compaction can be enabled via RunConfig(session_settings=SessionSettings(compaction=...))