Skip to main content
Model provider endpoints can be busy places; sometimes an LLM request will time out. When that happens, dartantic will invoke its built-in retry logic to handle temporary rate limits and failures.

​
Zero Configuration

The HTTP 429 “Too Many Requests” client error response status code indicates that the client has sent too many requests within a given time period. This mechanism, known as “rate limiting,” is employed by servers to prevent overload, ensure fair resource usage, and mitigate potential abuse or denial-of-service attacks. When dartantic gets a 429 response, it will retry automatically without you having to do anything: 
final agent = Agent('openai');

// Automatically retries on 429 errors
final result = await agent.send('Hello!');

​
Timing Strategy

Dartantic will use an incremental backoff algorithm along with a random “jitter” to see if it can get a response to go through. 
Attempt 1: Immediate
Attempt 2: 1s + jitter
Attempt 3: 2s + jitter  
Attempt 4: 4s + jitter

​
Next Steps