Usage & limits

Usage logging

Every request through api.mantler.ai is logged with:

• Timestamp
• API key identifier (not the key itself)
• Model and mantle ID
• Token counts (prompt, completion, total)
• Latency

Usage is visible on the Usage page in the Mantler web app, broken down by key and model.

---

Rate limits

Rate limits are applied per API key. Default limits are set at the org level.

When creating or editing a key, you can set lower per-key limits:

• Requests per minute (RPM)
• Tokens per minute (TPM)

When a limit is exceeded, the API returns 429 Too Many Requests. The response includes a Retry-After header indicating how many seconds to wait.

---

Token counting

Token counts in responses are approximate and depend on the runtime. For Ollama and llama.cpp backed models, the values reported in usage.prompt_tokens and usage.completion_tokens are the counts returned by the runtime.

---

Coming soon

Per-key spend limits and budget alerts are planned. Check the changelog for updates.