Chat that can work
Draft emails, summarize reports, write code, organize notes, and keep the current task in context.
Windows AI workspace
One clean local AI workspace.
Boolean brings chat, an embedded browser, notepad, local models, and optional cloud tokens into a simple desktop workspace built by saz3 Labs.
Built for private work first.
Use Boolean locally with GGUF models. Sign in only when you want optional cloud tokens or online providers.
Browser beside chat
Open pages, read visible content, send screenshots to notes, and keep research inside the same workspace.
Notepad included
Keep snippets, screenshots, drafts, and selected chat text in a clean paper-style notepad.
Local models
Run supported GGUF models on your PC with no required account for standard local use.
Optional cloud
Google Sign-In, Boolean Cloud tokens, or your own provider API key are available when you want online models.
User controlled
Local mode keeps prompts and files on the computer. Risky actions can require approval.
Offline models
Pick a model, download once, use it locally.
Boolean can help users choose a home-friendly model, download it into the right folder, and switch when the task changes.
Recommended for local
Qwen2.5-3B
Fast mode for most home PCs
Selected
Qwen2.5-7B
Daily chat and agent work
Download
Qwen2.5-VL-7B
Vision, screenshots, images
Download
Downloading Qwen2.5-7B
Saved to ~/.saz/models
Works offline after download
Use your own key
Connect leading AI providers when you want them.
Boolean does not lock you into one model. Stay local, sign in to Boolean Cloud, or add your own API key for supported online providers.
Default
Local models
Run GGUF models on your PC for private work and offline use.
Your key
Bring leading AI providers
Add a supported API key when you want a specific cloud model or account.
OpenAI
GLM
Z.ai
Claude
Boolean Cloud
Sign in instead
Use included cloud tokens without managing a separate model provider key.
API keys are stored on the user's computer. Boolean Cloud is separate and optional.
Compare
Boolean is built for local-first desktop work.
Other tools are great at their lane. Boolean combines the pieces people usually juggle: local models, chat, browser, notes, and optional cloud access.
Platform
Best for
Local models
Browser + notes
Cloud option
Booleanby saz3 Labs
One Windows workspace for asking, writing, searching, coding, and planning.
Built inGGUF downloads and local mode
Built inbrowser, notepad, screenshots
OptionalBoolean Cloud or your own key
Ollamalocal model runner
Running local models behind an API or terminal workflow.
Strongmodel serving focus
Separateneeds another app or UI
Separatebring another service
Cloud chat appsChatGPT / Claude style
High-quality cloud reasoning, writing, and general chat.
Noruns through cloud accounts
Variesusually web-first
Coresubscription or usage plan
Coding assistantsCopilot / IDE tools
Coding inside an editor or developer workflow.
Variesoften cloud-first
Separatenot a full workspace
Commonaccount-based access
App examples
More than a chat box.
Browser
Research without leaving the app
Keep the browser beside chat, then send page text or screenshots back to Boolean.
Date
Save drafts and screenshots
Use the built-in notepad for emails, snippets, visuals, and work you want to keep.
> create a small API
checking files...
ready
Local coding workflows
Let Boolean help with PowerShell, files, projects, and code while you stay in control.
Local mode
Your work stays on your PC.
Boolean ships as a Windows app with a built-in local backend and llama.cpp engine. No Node.js, Python, .NET runtime, or Ollama is required for the standard installed app.
Boolean Cloud
Sign in only when you want cloud tokens.
New Google Sign-In users may receive 100,000 cloud tokens for the first 30 days with a 10,000 token daily cap. The free tier currently targets GLM-4.7-Flash through Cloudflare Workers AI.
Requirements
WindowsWindows 10 or 11, x64 PC, WebView2 Runtime.
Cloud mode4 GB RAM, internet connection, Boolean Cloud account or provider API key.
Local LLM8 GB RAM recommended, 2 GB free disk space, AVX2 preferred.