Pixelfit — Image Sizing, Density & Weight Budgets

Pixelfit — Image Sizing, Density & Weight Budgets — a premium, 100% offline digital tool by Aceplore. Runs entirely in your browser: no accounts, no subscriptions, your data never leaves your device.

A private, offline planner for image sizing — the intrinsic pixels a slot really needs at each density, why 2x costs four times and not twice, and what your images spend of a page budget — 31 tools in one, with savings taken from bytes you measured rather than a rule of thumb.

31 built-in tools in one app. An image library with intrinsic sizes and the recorded bytes of each format you ship, sorted by weight; slots describing where an image appears and at what CSS size; a device set with its density and its share; a density view giving the intrinsic pixels a CSS size demands at each DPR and a required-size view naming the maximum across the whole device set; a retina-cost view that shows AREA rather than width driving the byte cost — 1,280 × 720 is 921,600 pixels against 230,400 at 1x, four times for twice the width — plus upscaling and downscaling lists naming images that are too small for their slot or far larger than anything needs; a srcset builder producing width descriptors and a sizes attribute from the slot itself; crops by ratio and focal point with the crop box drawn, the discarded share stated and any CLAMPING reported rather than silently applied; formats with their relative weight and a savings view computed from the bytes you recorded — never a generic percentage the app invented; a weight table, a page budget with an over-or-under verdict, a per-image budget, and lazy and hero views separating what can be deferred from what cannot; plus CSV exports, a JSON backup, and a Method page holding every formula with a worked example and the kilobyte convention in force.

Required size. Every device’s DPR applied to the slot’s CSS width, with the largest named as the one that governs. A DPR ceiling is a setting, so you can decide where 3x stops being worth it rather than shipping it by default.

Crop box. A 16:9 crop of a 4000 × 3000 source is 4000 × 2250, discarding 750 of height — 25% of the image. A focal point at 0.35 would start the box at −75, so it clamps to 0, and Pixelfit SAYS it clamped rather than moving your focal point behind your back.

Budget check. 1,460 KB of images against a 1,200 KB budget is 260 KB over — 21.666667% of the budget — with the responsible images listed largest first. The kilobyte convention (1,024 or 1,000 bytes) is a setting and is printed beside the figure.

What's included:

• Pixelfit.html
• Quick-Start-Guide.html
• How-To-Use-Guide.html
• README.txt
• LICENSE.txt

Includes a personalized welcome wizard, an interactive guided tour, a ⌘K command palette, light/dark themes with 8 accents, and one-click JSON backup & restore.

Pixelfit is a specification, costing and record-keeping tool for design and production work. Every figure it reports is computed by a stated formula from the sizes, prices, counts and rates YOU enter, and every figure carries its UNIT, because px, pt, mm, bytes, milliseconds and dpi are different quantities and confusing them is the classic failure in this field. Where a geometry has to close — a grid against its container, a proration, a coverage matrix — it is reconciled rather than rounded, and a disagreement is reported instead of absorbed. Where a rule comes from a published standard or an industry convention, the standard is named and its parameters are yours to change; where a figure rests on something you assumed rather than measured, it is labelled an assumption at the point of use. Nothing here is a guarantee of a passed press check, a performance score, an approved brand or an available name, and none of it is legal advice — trademark, licensing and accessibility obligations differ by country and by contract, and availability and registration must be confirmed with the registries and rights holders themselves. You are solely responsible for what you send to press, ship to a browser or put your name to.