Local by design
Processing happens in the Bun runtime on your device. Images never cross a web API or an account boundary.
Local-first image workflows
Optimize, transform, and organize image batches with a calm desktop workflow or a predictable command line. Your files stay on your machine.
Example command: rastry run ./public --preset web. The dry run plans four files and completes without changing them.
A smaller, safer loop
Rastry is built for developers, designers, and small teams who want repeatable image work without handing assets to a service.
Processing happens in the Bun runtime on your device. Images never cross a web API or an account boundary.
Express a workflow as a versioned pipeline or a readable preset, then use the same rules from CLI or desktop.
Every run starts with a plan. Outputs get a derived name, originals are protected, and conflicts are visible before writing.
Current foundation
Rastry currently handles the formats and transformations that make up the everyday web-asset loop.
Read the tradeoffs and use each format deliberately.
Explore formatsShape dimensions, remove transparent borders, and prepare assets for their destination.
Explore operationsStart with shipped recipes that map to real delivery contexts.
Explore presetsChoose your entry point
The CLI and Desktop v0.2 use the same pipeline contract and safety rules. Choose the surface that fits the task.
Plan a one-off conversion, run a preset over a folder, or capture a JSON summary for automation.
rastry photo.png --to webp --quality 82 --dry-runSelect files or folders, configure operations, review the plan, then execute only after the output paths look right.
Trust is a feature
Rastry favors a visible, deterministic plan over a surprising shortcut.
-rastry suffix.Maturity, made visible
Rastry is foundation-stage software. The roadmap is a promise about sequence, not a claim that every item is shipped.
Core, Bun.Image adapter, safe CLI, formats, operations, batches, and dry-run.
Electrobun + React workflow with typed RPC, plan preview, progress, and summary.
Local presets, import/export, execution history, and recipe documentation.
Explicit watch mode, installers, updates, and performance work.
Open source, local first
Read the contribution guide, try a real web-asset workflow, and report what felt unclear. The project is deliberately early enough for useful feedback to change the shape.