specificationVersion: "0.1" id: filebase-rate-limits name: Filebase API Rate Limits description: >- Filebase enforces rate limits per API surface to ensure fair use and platform stability. The S3-Compatible API and Platform API share the same 500 requests-per-second limit per account. The IPFS Pinning Service API is separately limited to 100 requests per second. The IPFS RPC API inherits the same token-scoped constraints as the Pinning Service API. Limits apply at the account level, not per bucket or per key. url: https://filebase.com/docs/platform-api/overview rateLimits: - api: Filebase S3-Compatible API description: >- Bucket and object operations via the S3 protocol at s3.filebase.io. Rate limit matches AWS S3 defaults to preserve compatibility with existing tooling and SDK expectations. limits: - scope: account period: second requests: 500 unit: requests/second notes: >- Per-account limit. Individual bucket prefix limits may apply for high-throughput multipart uploads per S3 best practices. - api: Filebase Platform API description: >- Account-level usage metrics and management endpoints at api.filebase.io. Limit is shared with the S3 API rate envelope. limits: - scope: account period: second requests: 500 unit: requests/second notes: >- Authenticated via HTTP Basic / Bearer token using access key pairs. - api: Filebase IPFS Pinning Service API description: >- IPFS pinning operations (list, add, get, replace, delete) at api.filebase.io/v1/ipfs/pins. Lower limit reflects pin queue management overhead. limits: - scope: account period: second requests: 100 unit: requests/second notes: >- Authenticated via per-bucket Bearer token. Async pin status polling should be implemented with exponential backoff to stay within this limit. - api: Filebase IPFS RPC API description: >- Core IPFS daemon operations at rpc.filebase.io. All endpoints are POST requests authenticated with bucket-specific Bearer tokens. limits: - scope: bucket period: second requests: 100 unit: requests/second notes: >- Tokens are bucket-scoped; limits apply per token/bucket combination. Heavy add/get operations may be subject to additional throughput constraints based on object size. headers: description: >- Filebase follows standard HTTP conventions for communicating rate limit state. Clients should handle 429 Too Many Requests responses with exponential backoff and respect Retry-After headers when present. statusCodes: - code: 429 meaning: Too Many Requests — rate limit exceeded; retry after backoff - code: 503 meaning: Service temporarily unavailable — check status.filebase.com bestPractices: - Implement exponential backoff with jitter on 429 and 503 responses. - Use multipart uploads for large objects to parallelize within limits. - Cache IPNS resolution results locally to avoid repeated RPC calls. - Poll pin status with increasing intervals (start at 5s, cap at 60s). - Use per-bucket tokens for IPFS APIs to isolate rate limit consumption.