Rate limits, best practices, and error handling

Rate limits

• Quote / build / cross-chain / submit-solana: 300 requests per minute per IP
• Chains, tokens, settings, record: 120 requests per minute per IP
• Responses include RateLimit-Remaining header

Common error codes

• RATE_LIMITED (429) — exponential backoff starting at 2s
• NO_LIVE_QUOTES (503) — pair may be illiquid; try smaller size or different tokens
• SAME_TOKEN (400) — tokenIn and tokenOut must differ
• VALIDATION_ERROR (400) — check Zod shapes in OpenAPI spec

Best practices

Cache GET /swap/tokens results for at least 60 seconds. Re-quote only when amount or slippage changes — do not quote on every block. Always pass taker at quote time on EVM meta routes for accurate gas estimates. Use X-Client-Id consistently so admin can attribute traffic.