Project description
TranslatedWe have several content sites and we're tired of manually pre-generating dozens of sizes for every image and endlessly storing them on disk. We want a proper image processing service that works on the fly: I point to a source (either an upload endpoint or existing storage we have), and it serves exactly the variant I requested, straight through a URL. Width, height, crop mode, quality and target format should be set via query parameters or a clean path scheme so frontend can request the right size for each device and layout without rebuilds or manual work.
The essence of the task is quality dynamic transformation plus delivery. On first request for a variant, the service renders it, puts it in edge cache and all subsequent requests serve from cache with proper cache headers and fast time to first byte. Modern formats are important: automatic delivery of WebP and AVIF depending on what the browser accepts, sensible compression and optional metadata stripping. We need smart cropping so portraits and product cards aren't cut off when scaled to square, and DPR support for retina screens. Separately explain how incorrect and oversized requests are rejected so the resizer can't be used as an open proxy.
To set scope clearly: this is not a media library or DAM, and not a universal CDN for arbitrary files. The focus is strictly on dynamic image transformation and fast delivery of results to the visitor. We'll provide the source storage and staging. Send similar projects, the proposed stack and ballpark numbers for throughput and latency you can guarantee.
— Resize, crop and format conversion via URL or path scheme
— Edge cache with proper headers and fast re-delivery
— Auto WebP/AVIF, quality control, metadata stripping
— Smart crop, DPR/retina support, request signing or validation
— Documentation and brief handoff of running pipeline
The essence of the task is quality dynamic transformation plus delivery. On first request for a variant, the service renders it, puts it in edge cache and all subsequent requests serve from cache with proper cache headers and fast time to first byte. Modern formats are important: automatic delivery of WebP and AVIF depending on what the browser accepts, sensible compression and optional metadata stripping. We need smart cropping so portraits and product cards aren't cut off when scaled to square, and DPR support for retina screens. Separately explain how incorrect and oversized requests are rejected so the resizer can't be used as an open proxy.
To set scope clearly: this is not a media library or DAM, and not a universal CDN for arbitrary files. The focus is strictly on dynamic image transformation and fast delivery of results to the visitor. We'll provide the source storage and staging. Send similar projects, the proposed stack and ballpark numbers for throughput and latency you can guarantee.
— Resize, crop and format conversion via URL or path scheme
— Edge cache with proper headers and fast re-delivery
— Auto WebP/AVIF, quality control, metadata stripping
— Smart crop, DPR/retina support, request signing or validation
— Documentation and brief handoff of running pipeline
Other / Miscellaneous Cash, Bank transfer
