Claude Code API Error 529 overloaded: causes and fixes

The error text

API Error: 529 {"type":"error","error":{"type":"overloaded_error","message":"Overloaded"}}

What it means

The API accepted your key and your request, but the model was overloaded and could not take it. Peaks are short and tied to a specific model, which is why switching model or waiting a minute usually works.

Step-by-step fix

1. 1. Retry

   Send the message again after a short pause. Claude Code keeps the conversation, so nothing is lost.

2. 2. Switch model

   Run /model and pick another Claude model (for example Sonnet instead of Opus) to finish the task, then switch back later.

3. 3. Switch channel

   On aicodewith, each model page shows channel health. If one channel is busy, choose another for your key.

4. 4. Check your endpoint

   Run /status and confirm the endpoint is api.aicodewith.ai. A wrong endpoint shows other errors, but it is worth ruling out.

   ~/.claude/settings.json

   {
     "env": {
       "ANTHROPIC_AUTH_TOKEN": "YOUR_API_KEY",
       "ANTHROPIC_API_KEY": "",
       "ANTHROPIC_BASE_URL": "https://api.aicodewith.ai"
     }
   }

How to avoid it

• •Use Sonnet or Haiku for routine work and save Opus for hard problems; busy peaks hit the largest models first.
• •Run fewer parallel subagents during peak hours.
• •For CI jobs, retry with backoff instead of failing the job on the first 529.

Are failed requests charged?

Failed requests (4xx/5xx, timeouts, dropped connections) are not charged — you're only billed for requests that complete successfully. If an upstream failure is automatically retried and succeeds, you're billed once.

Still failing?

• •Check the model's page on aicodewith: it shows recent health for each channel.
• •Look up the request in your dashboard usage records to see its model, status and cost.
• •Email support@aicodewith.com with the time of the error, the model and the error text (never your full API key).

Model status and pricesContact supportDocs

FAQ