Python Async With AI

async/await lets one program wait on many things at once. Perfect for HTTP calls and LLM APIs. Let AI help you avoid the common traps.

45 min · Reviewed 2026

Concurrency, Not Parallelism

async is about waiting efficiently, not running faster on many CPUs. One event loop juggles hundreds of network calls by pausing each one while it waits.

import asyncio
import httpx

async def fetch(client: httpx.AsyncClient, url: str) -> dict:
    r = await client.get(url, timeout=10)
    r.raise_for_status()
    return r.json()

async def main(urls: list[str]) -> list[dict]:
    async with httpx.AsyncClient() as client:
        return await asyncio.gather(*(fetch(client, u) for u in urls))

if __name__ == "__main__":
    data = asyncio.run(main([
        "https://httpbin.org/json",
        "https://httpbin.org/uuid",
    ]))
    print(data)httpx.AsyncClient plus asyncio.gather fetches many URLs concurrently without threads.

Bounded concurrency with a semaphore

async def fetch_limited(sem: asyncio.Semaphore, client: httpx.AsyncClient, url: str):
    async with sem:
        r = await client.get(url)
        return r.json()

async def many(urls: list[str]) -> list[dict]:
    sem = asyncio.Semaphore(10)  # at most 10 in flight
    async with httpx.AsyncClient() as client:
        return await asyncio.gather(*(fetch_limited(sem, client, u) for u in urls))Semaphores cap how many coroutines run at once. Essential when hitting rate-limited APIs.

Understanding "Python Async With AI" in practice: AI-assisted coding shifts work from syntax recall to design thinking — models handle boilerplate so you focus on architecture. Async/await lets one program wait on many things at once. Perfect for HTTP calls and LLM APIs. Let AI help you avoid the common traps — and knowing how to apply this gives you a concrete advantage.

Apply asyncio in your ai-coding workflow to get better results
Apply await in your ai-coding workflow to get better results
Apply gather in your ai-coding workflow to get better results
Apply httpx in your ai-coding workflow to get better results

Use AI to generate unit tests for an existing function
Ask AI to refactor a messy function and explain the changes
Have AI suggest a code review for a recent pull request

The big idea: async is the right tool when you wait more than you compute. Bound concurrency, stay inside one run() call, and never block the loop.

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-progx-python-async-creators

What is the primary performance benefit of using async/await in Python?
1. It automatically speeds up network requests by compressing data
2. It eliminates the need for any waiting in asynchronous code
3. It allows multiple CPU-heavy tasks to run on different processor cores simultaneously
4. It enables efficient handling of I/O-bound operations by minimizing idle waiting time
Which code snippet correctly pauses for 3 seconds within an async function?
1. await asyncio.sleep(3)
2. time.sleep(3)
3. asyncio.sleep(3)
4. await time.sleep(3)
What is the purpose of a semaphore in asynchronous Python code?
1. To measure how fast asynchronous tasks complete
2. To store results from multiple async tasks in order
3. To automatically convert synchronous functions to asynchronous ones
4. To limit the number of concurrent operations happening at once
What happens when you nest one asyncio.run() call inside another?
1. The nested call is silently ignored and only the outer run executes
2. Both runs execute sequentially without issues
3. The inner run creates a new event loop automatically
4. A RuntimeError is raised indicating nested run calls are not allowed
Which statement best describes what an event loop does?
1. It runs code on multiple CPU cores in parallel
2. It prevents any errors from occurring in async code
3. It automatically optimizes slow database queries
4. It switches between waiting tasks rather than leaving CPUs idle
You need to make 100 API calls. What does async allow you to do that synchronous code cannot?
1. Execute the requests across 100 different CPU cores
2. Pause some requests while waiting for others, maximizing throughput without parallelism
3. Complete all requests instantly without any network latency
4. Send all 100 requests simultaneously with no delay between them
When is async programming the right tool for a task?
1. When the task involves heavy mathematical computations like matrix multiplication
2. When the task must complete as fast as possible regardless of resource usage
3. When the task spends more time waiting for external resources than processing data
4. When the task requires running multiple separate programs at once
What does asyncio.gather() do when passed multiple coroutines?
1. Cancels any that take longer than 10 seconds automatically
2. Executes them one after another in the order provided
3. Requires a semaphore to be passed as the first argument
4. Runs them concurrently and returns all results when complete
What is a common pitfall when calling external APIs in a loop using async?
1. Forgetting to use the await keyword before the API call
2. All async API calls will fail due to missing API keys
3. Launching too many concurrent requests and overwhelming the API or hitting rate limits
4. Using time.sleep() instead of asyncio.sleep() which actually works fine
What does 'bounded concurrency' mean in async programming?
1. Ensuring all tasks complete before the program exits
2. Running exactly two tasks at a time no matter what
3. Limiting how many async tasks can run simultaneously to a specific number
4. Using only one semaphore in the entire program
Why is calling asyncio.run() only once at program startup recommended?
1. Multiple calls waste memory and slow down the computer
2. The function automatically closes all network connections after first use
3. Nesting calls causes RuntimeError and creates confusing event loop states
4. asyncio.run() can only be used in the main function, not in any other function
Which scenario best demonstrates async's strength over synchronous code?
1. Calculating prime numbers up to 1 million
2. Reading a small config file from disk
3. Downloading 50 web pages where each request takes 1 second
4. Sorting a list of 1 million numbers
What is wrong with this code: async def fetch(): return await requests.get(url)?
1. async def is not allowed to return values
2. await is unnecessary when returning from an async function
3. The syntax is incorrect and will cause a parse error
4. requests.get() is synchronous and will block the event loop when called directly
How does an event loop handle 500 concurrent HTTP requests?
1. It switches between requests while they wait for responses, handling all with one thread
2. It processes them in batches of 10 using built-in batch processing
3. It processes all requests simultaneously on a multi-core CPU
4. It creates 500 separate threads to handle each request
What happens if you use await on a coroutine but forget to await a different one in your async function?
1. The event loop will automatically await it later
2. It causes undefined behavior - the coroutine may or may not run
3. The program will crash immediately with a syntax error
4. The unawaited coroutine never executes

← Back to interactive lesson

Tendril · Creators · AI-Assisted Coding

Python Async With AI

async/await lets one program wait on many things at once. Perfect for HTTP calls and LLM APIs. Let AI help you avoid the common traps.

45 min · Reviewed 2026

Concurrency, Not Parallelism

async is about waiting efficiently, not running faster on many CPUs. One event loop juggles hundreds of network calls by pausing each one while it waits.

import asyncio
import httpx

async def fetch(client: httpx.AsyncClient, url: str) -> dict:
    r = await client.get(url, timeout=10)
    r.raise_for_status()
    return r.json()

async def main(urls: list[str]) -> list[dict]:
    async with httpx.AsyncClient() as client:
        return await asyncio.gather(*(fetch(client, u) for u in urls))

if __name__ == "__main__":
    data = asyncio.run(main([
        "https://httpbin.org/json",
        "https://httpbin.org/uuid",
    ]))
    print(data)httpx.AsyncClient plus asyncio.gather fetches many URLs concurrently without threads.

Bounded concurrency with a semaphore

async def fetch_limited(sem: asyncio.Semaphore, client: httpx.AsyncClient, url: str):
    async with sem:
        r = await client.get(url)
        return r.json()

async def many(urls: list[str]) -> list[dict]:
    sem = asyncio.Semaphore(10)  # at most 10 in flight
    async with httpx.AsyncClient() as client:
        return await asyncio.gather(*(fetch_limited(sem, client, u) for u in urls))Semaphores cap how many coroutines run at once. Essential when hitting rate-limited APIs.

Apply asyncio in your ai-coding workflow to get better results
Apply await in your ai-coding workflow to get better results
Apply gather in your ai-coding workflow to get better results
Apply httpx in your ai-coding workflow to get better results

Use AI to generate unit tests for an existing function
Ask AI to refactor a messy function and explain the changes
Have AI suggest a code review for a recent pull request

The big idea: async is the right tool when you wait more than you compute. Bound concurrency, stay inside one run() call, and never block the loop.

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-progx-python-async-creators

What is the primary performance benefit of using async/await in Python?
1. It automatically speeds up network requests by compressing data
2. It eliminates the need for any waiting in asynchronous code
3. It allows multiple CPU-heavy tasks to run on different processor cores simultaneously
4. It enables efficient handling of I/O-bound operations by minimizing idle waiting time
Which code snippet correctly pauses for 3 seconds within an async function?
1. await asyncio.sleep(3)
2. time.sleep(3)
3. asyncio.sleep(3)
4. await time.sleep(3)
What is the purpose of a semaphore in asynchronous Python code?
1. To measure how fast asynchronous tasks complete
2. To store results from multiple async tasks in order
3. To automatically convert synchronous functions to asynchronous ones
4. To limit the number of concurrent operations happening at once
What happens when you nest one asyncio.run() call inside another?
1. The nested call is silently ignored and only the outer run executes
2. Both runs execute sequentially without issues
3. The inner run creates a new event loop automatically
4. A RuntimeError is raised indicating nested run calls are not allowed
Which statement best describes what an event loop does?
1. It runs code on multiple CPU cores in parallel
2. It prevents any errors from occurring in async code
3. It automatically optimizes slow database queries
4. It switches between waiting tasks rather than leaving CPUs idle
You need to make 100 API calls. What does async allow you to do that synchronous code cannot?
1. Execute the requests across 100 different CPU cores
2. Pause some requests while waiting for others, maximizing throughput without parallelism
3. Complete all requests instantly without any network latency
4. Send all 100 requests simultaneously with no delay between them
When is async programming the right tool for a task?
1. When the task involves heavy mathematical computations like matrix multiplication
2. When the task must complete as fast as possible regardless of resource usage
3. When the task spends more time waiting for external resources than processing data
4. When the task requires running multiple separate programs at once
What does asyncio.gather() do when passed multiple coroutines?
1. Cancels any that take longer than 10 seconds automatically
2. Executes them one after another in the order provided
3. Requires a semaphore to be passed as the first argument
4. Runs them concurrently and returns all results when complete
What is a common pitfall when calling external APIs in a loop using async?
1. Forgetting to use the await keyword before the API call
2. All async API calls will fail due to missing API keys
3. Launching too many concurrent requests and overwhelming the API or hitting rate limits
4. Using time.sleep() instead of asyncio.sleep() which actually works fine
What does 'bounded concurrency' mean in async programming?
1. Ensuring all tasks complete before the program exits
2. Running exactly two tasks at a time no matter what
3. Limiting how many async tasks can run simultaneously to a specific number
4. Using only one semaphore in the entire program
Why is calling asyncio.run() only once at program startup recommended?
1. Multiple calls waste memory and slow down the computer
2. The function automatically closes all network connections after first use
3. Nesting calls causes RuntimeError and creates confusing event loop states
4. asyncio.run() can only be used in the main function, not in any other function
Which scenario best demonstrates async's strength over synchronous code?
1. Calculating prime numbers up to 1 million
2. Reading a small config file from disk
3. Downloading 50 web pages where each request takes 1 second
4. Sorting a list of 1 million numbers
What is wrong with this code: async def fetch(): return await requests.get(url)?
1. async def is not allowed to return values
2. await is unnecessary when returning from an async function
3. The syntax is incorrect and will cause a parse error
4. requests.get() is synchronous and will block the event loop when called directly
How does an event loop handle 500 concurrent HTTP requests?
1. It switches between requests while they wait for responses, handling all with one thread
2. It processes them in batches of 10 using built-in batch processing
3. It processes all requests simultaneously on a multi-core CPU
4. It creates 500 separate threads to handle each request
What happens if you use await on a coroutine but forget to await a different one in your async function?
1. The event loop will automatically await it later
2. It causes undefined behavior - the coroutine may or may not run
3. The program will crash immediately with a syntax error
4. The unawaited coroutine never executes

← Back to interactive lesson