Private by design · powered by WebLLM

Edit with total privacy using alocal language model.

Refine essays, fiction, and personal writing directly in your browser. Your text stays on your device, your model remains your choice, and every change is yours to review.

No accountNo text uploadRuns after model download

Focused tools, fewer distractions

Built around the edit

Write and compare

Stream an edit, inspect the original beside the result, and keep complete run history locally.

Process long work

Apply a proven instruction paragraph by paragraph, pause safely, and review every change before export.

Return to decisions

Restore prompts, source text, output, model settings, and conversations from this browser.

What “local” means here

The model downloads. Your writing does not upload.

WebLLM fetches model files when you first select a model, then performs inference through WebGPU. AIditorial stores runs and extracted document text in your browser's IndexedDB without a remote application server.