Guest workspace · nothing to sign up for

Settings

Your account, plan, appearance, privacy choices and the AI assistant — all on one page.

Account

You are working as a guest. Every tool is available without an account.

Keep favorites and workflows per account
Unlock Pro limits once billing opens
Still nothing uploaded — documents stay on your device

Plan & billing

Limits below are what the tools enforce on this plan.

Free

File size50 MB

Files per batch20

Pages per task200

Saved workflows3

AdvertisementsShown

Cloud AIBring your own key

See Pro — $4/month billed yearlyPaid plans are in preview; no payments are collected yet.

Appearance

Theme is remembered on this device. Reduced-motion settings from your system are respected automatically.

Privacy & data

What stays on your device, what you have allowed, and how to clear it.

Document processing: On your device
Document uploads: None
Document conversions: On your device
AI assistant: Device-first; cloud engines only with your consent, text only
Third-party advertising: Not enabled in preview

Anonymous page statisticsCounts page views and active visitors only. No document data, no cross-site tracking, honours Do Not Track.

Off

Saved documentsLocal storage is unavailable in this browser.

Workflows and favorites0 workflows · 0 favorites, kept in this browser.

Read the full privacy details

AI assistant

Choose where documents are understood. On-device engines are free and private; cloud engines are included in paid plans or billed to your own provider key.

Preferred engine

Automatic picks the cheapest private engine that is ready on this device, and falls back to cloud when you are entitled to it.

On-device model

Runs on your GPU through WebGPU. Downloaded once, cached by the browser, works offline.

Qwen3 1.7BRecommended. Strong summaries and Q&A in 100+ languages. · 1096.7 MB download · ~2.0 GB VRAMLlama 3.2 1BLightest. Good for laptops with small GPUs. · 839.2 MB download · ~0.9 GB VRAMQwen3 4BBest quality. Needs a desktop-class GPU (4 GB+ VRAM). · 2479.6 MB download · ~3.4 GB VRAMLlama 3.2 3BBalanced quality and speed for English documents. · 2002.7 MB download · ~2.2 GB VRAM

Your own provider key

Bring a key from any OpenAI-compatible provider or Anthropic. It is stored only in this browser and sent straight to the provider — never to our servers.

Provider Very low cost; generous free tier in AI Studio.Model

Base URLAPI key

Self-hosted endpoints (Ollama, vLLM, LM Studio) must allow browser requests from this site (CORS).

Keyboard shortcuts

Use Ctrl instead of ⌘ on Windows and Linux.

⌘ K: Search tools from anywhere
Esc: Close menus and dialogs
⌘ V: Paste a file into a tool
Shift + click: Select a range of pages
← →: Move a page while reordering
Enter / Space: Open or toggle the focused card