Rate limit 429 api proxy

Error 429 Too Many Requests

Repeated 429s usually mean the upstream account pool is rate limited, cooling down, or out of usable quota. Ask the provider to switch account, route, or model first, then run a minimal request to rule out local concurrency, balance, or retry-loop issues.

Snapshot

Error: exceeded retry limit, last status: 429 Too Many Requests, request id: <request-id>. With an AI API proxy, repeated 429s usually mean the upstream account pool is rate limited, cooling down, or out of usable quota. Ask the provider to switch account, route, or model first.

#error 429 #429 too many requests #openai api 429 #claude code 429 #codex 429 #exceeded retry limit 429

What Error 429 Means

Error 429 means the server accepted the general shape of the request, but refused to process it right now because a quota or rate limit was hit.

With an AI API proxy, look at the provider’s upstream account pool before retrying the same API key. If that pool is cooling down or exhausted, repeated retries only amplify the 429.

Common Causes

  1. The proxy provider’s upstream account pool hit rate limits, entered cooldown, or no longer has usable quota.
  2. The current model route has no active billing capacity, hit a model-specific limit, or has no healthy upstream account available.
  3. The API proxy provider is rate limiting your key even if another account pool or the upstream official API still has capacity.
  4. Claude Code, Codex, Cursor, or another tool may retry failed calls and quickly turn one failed operation into many requests.
  5. A high-context prompt, streaming task, batch job, or coding-agent loop may consume tokens faster than expected.

How To Diagnose 429 For AI API Proxies

Start with the smallest possible request against the same Base URL, API key, and model. If that request also returns 429, the problem is likely quota, billing, provider limits, or proxy-side throttling.

If the minimal request works but a coding tool fails, reduce agent concurrency and look for retry loops. Coding agents can rapidly send many small requests while reading files, calling tools, or recovering from a failed step.

How To Fix 429 For AI API Proxies

  1. Wait before retrying. If the response includes a Retry-After header, use that delay instead of immediately sending the same request again.
  2. Check the provider dashboard for balance, billing status, daily quota, request limits, and model-specific restrictions.
  3. Confirm the concurrency limit for your current API proxy plan or API key, then reduce concurrency, lower max output tokens, shorten prompts, and avoid running multiple coding-agent sessions against the same API key.
  4. If the current Claude Code, Codex, or Cursor session keeps returning 429 while the minimal request works, stop the current task, wait briefly, and retry in a new chat or fresh session.
  5. If 429 keeps happening, contact the API proxy provider’s support team with the error time, model name, Base URL, client tool, and full error message so they can check account-pool capacity, quota, or route-level throttling.

Related errors

Claude Code / Codex 503 No available accounts Do not rotate keys first. This error usually points to an unavailable account pool, provider, group, or model route. Test a minimal request, confirm the model and group, then decide whether to wait, reduce context, change model, or switch provider. Codex: Selected model is at capacity. Error: Selected model is at capacity. Please try a different model. Treat it as model-level capacity, admission, or streaming interruption first. It can happen with ChatGPT-authenticated Codex CLI, proxy account pools, or Sub2API upstream overloaded passthrough. Codex warning: Your conversation has multiple possible cybersecurity risk flags Warning: Your conversation has multiple possible cybersecurity risk flags. Treat it as an OpenAI/Codex safety-routing signal first, then check whether an API proxy, sensitive keywords, long code context, or opaque model routing is making it worse. Error 401 Unauthorized for AI API proxies A 401 is an authentication failure. Check which key is loaded, which Base URL receives the request, whether the header format is correct, and whether the key is still active for that provider.

Related guides

Quickstart: connect Claude Code, Codex, or Cursor to an AI API proxy Before connecting Claude Code, Codex, or Cursor to an AI API proxy, confirm the tool, protocol compatibility, Base URL, API key, and model name. Codex OpenAI-compatible proxy Base URL setup guide Install the OpenAI Codex CLI and configure API keys, Base URL, config.toml, and provider switching for an OpenAI-compatible proxy. Claude Code API proxy Base URL setup guide Configure Claude Code with an API proxy using ANTHROPIC_BASE_URL, ANTHROPIC_AUTH_TOKEN, and ANTHROPIC_API_KEY.

Related topics

OpenAI API Proxy Providers Compare OpenAI API proxy and OpenAI-compatible providers by Base URL support, models, pricing, payment methods, and public verification signals. Claude API Proxy Providers Compare Claude API proxy and Anthropic-compatible providers by models, Base URL setup, pricing notes, payment methods, and verification signals. AI API Directory Browse an AI API directory of OpenAI-compatible, Claude-compatible, and multi-model proxy providers with pricing, payment, model, and verification fields.

FAQ

Common questions

Is Error 429 always a provider outage?

No. It is more often a rate-limit, quota, billing, or retry-loop issue. Check account balance, request limits, and Retry-After before assuming the provider is down.

Why does Error 429 happen more often in Codex or Claude Code?

Coding agents can make many API calls while reading files, planning edits, running tools, and retrying failures. One interactive task may consume far more requests and tokens than a simple chat message.

Should I immediately switch API proxy providers after a 429?

Do not switch proxy providers first. Ask the current proxy provider to switch to a healthy upstream account, route, or model. Switching providers makes sense only when multiple account pools or routes keep returning 429 and the minimal request also fails.