RateLimitError
deepseek.RateLimitError (HTTP 429)
Stack trace
deepseek.RateLimitError: Error code: 429 - {'error': {'message': 'Rate limit exceeded', 'type': 'requests', 'code': 'rate_limit_exceeded'}} Why it happens
DeepSeek enforces strict rate limits to protect service stability. When your application sends more requests than allowed per minute or hour, the API responds with a 429 RateLimitError to throttle usage.
Detection
Monitor API response codes and catch deepseek.RateLimitError exceptions to detect when request limits are exceeded before your app crashes.
Causes & fixes
Sending too many requests in a short period exceeding DeepSeek's rate limit quota
Implement request throttling or exponential backoff retry logic to reduce request frequency and respect DeepSeek's rate limits.
Not using API key or using an invalid/expired API key causing default low rate limits
Ensure your API key is valid, correctly set in environment variables, and passed in all DeepSeek client requests.
Parallel or concurrent requests exceeding the allowed concurrency limits
Limit concurrent API calls by queuing requests or using a concurrency control mechanism like asyncio.Semaphore.
Code: broken vs fixed
from deepseek import DeepSeek
client = DeepSeek(api_key='hardcoded_key') # bad: hardcoded key
# This call may raise RateLimitError if too many requests are sent
response = client.search('example query') # triggers RateLimitError
print(response)import os
from deepseek import DeepSeek, RateLimitError
client = DeepSeek(api_key=os.environ['DEEPSEEK_API_KEY']) # fixed: use env var
try:
response = client.search('example query') # safe call with error handling
print(response)
except RateLimitError:
print('Rate limit exceeded, retry later or backoff')Workaround
Catch RateLimitError exceptions and implement a manual delay (e.g., time.sleep) before retrying the request to avoid immediate repeated failures.
Prevention
Use client-side rate limiting and exponential backoff retry strategies, monitor usage quotas, and ensure valid API keys to prevent hitting DeepSeek rate limits.