> ## Documentation Index
> Fetch the complete documentation index at: https://reagent-ai.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Async Support

> Use reagent-flow sessions with async/await.

## Overview

Sessions work as both sync and async context managers. The API is identical — use `async with` instead of `with`.

## Usage

```python theme={null}
import reagent_flow

async def test_async_agent():
    async with reagent_flow.session("async_flow", trace_dir=".reagent") as s:
        s.log_llm_call(
            tool_calls=[{"name": "search", "arguments": {"q": "test"}}],
        )
        s.log_tool_result("search", result={"found": True})

    s.assert_called("search")
```

## Thread safety

Sessions use Python's `contextvars`, which are natively async-safe. Each async task sees its own active session — no locking required.

```python theme={null}
import asyncio
import reagent_flow

async def run_agent(name: str):
    async with reagent_flow.session(name) as s:
        s.log_llm_call(tool_calls=[{"name": "work", "arguments": {}}])
        s.log_tool_result("work", result={"done": True})
    s.assert_called("work")

async def test_parallel_agents():
    await asyncio.gather(
        run_agent("agent-1"),
        run_agent("agent-2"),
        run_agent("agent-3"),
    )
```

Each session is isolated — `agent-1`'s trace never leaks into `agent-2`.