Bunny Cdn

Bunny.net (formerly BunnyCDN) positions itself as the affordable, simple CDN — and the API reflects that ethos. For agents, the REST API covers pull zone management (CDN distribution), storage zone operations (file storage), DNS zone management, and purge operations. The data model is straightforward: pull zones distribute content, storage zones host files, and the two can be linked. Edge scripting enables custom request/response logic. Stream API handles video delivery and encoding. Pricing is per-GB with no minimum commitments, making it cost-effective for variable agent workloads. The main trade-off versus Cloudflare or Fastly: fewer enterprise features (no WAF, limited DDoS, no Workers-equivalent ecosystem), but adequate CDN capability at a fraction of the cost.

REST API at api.bunny.net with clean endpoints for pull zones, storage zones, DNS zones, and statistics. Pull zone CRUD is straightforward: create, configure caching rules, set up custom hostnames, and purge. Storage zones provide S3-like file operations (upload, download, list, delete) via a separate storage API endpoint. Purge operations support URL-specific and full zone purging. Statistics API provides bandwidth, request, and cache hit rate data with hourly/daily granularity. DNS management supports standard record types. The API surface is intentionally small — covering core CDN operations without enterprise complexity. Pagination uses simple limit/offset. Response payloads are consistently JSON. No GraphQL or batch operation support.

API errors return JSON with Message field. HTTP status codes follow conventions: 401 for auth, 400 for bad requests, 404 for not found. Rate limits exist but aren't publicly documented with specific thresholds — the documentation says to implement reasonable backoff. Purge propagation is fast (typically under 5 seconds globally). Storage API file upload errors include size and path validation messages. The CDN infrastructure uses 100+ global PoPs. Cache behavior is configurable per pull zone (TTL, query string handling, vary headers). Edge scripting errors surface in the pull zone's error log. The main reliability consideration: Bunny.net is a smaller company than Cloudflare or Fastly, meaning SLA enforcement and enterprise support responsiveness differ. For cost-sensitive agent workloads, the reliability is adequate.

Single API key (AccessKey) passed via AccessKey header for all management API operations. Storage zones use a separate storage API key for file operations. No OAuth, no scoped permissions, no temporary credentials. The API key grants full account access — there's no way to create read-only or zone-limited keys. For agents managing a single account, this simplicity works well. For multi-tenant agents or environments requiring least-privilege access, the lack of scoping is a real limitation. TLS is enforced. No IP-based restrictions. The storage API key per-zone provides some isolation — agents can have per-zone file access without management API access.

Documentation at docs.bunny.net is clean and well-organized for the API's scope. API reference covers endpoints with parameters and response examples. Getting started guides are step-by-step. The documentation honestly represents the product's scope — it doesn't over-promise enterprise features. Storage API documentation is clear for file management operations. Client libraries: community libraries exist for various languages but no official SDKs. The documentation's main gap: advanced configuration patterns (edge scripting, custom cache keys, token authentication) have thinner documentation than core CDN operations. Dashboard UX is simple and fast. For agents, the documentation-to-integration path is short — the API surface is small enough to learn in an hour.