What’s new in Sato.

Companions & battery

More companions, less background work

Browse thousands of community pets from Petdex while Sato's new power modes keep the companion playful without continuously taxing your battery.

• Search and filter the Petdex catalog, preview each pet, and keep your selected companion available offline.
• After one minute without using Sato, the companion dozes in a stationary 8–10 FPS animation instead of patrolling.
• After 15 minutes of Mac inactivity, Sato suspends its overlays and animation until you return.
• Screen sleep, session lock, and permission changes now preserve in-progress work while pausing visual activity.

Sato Local

Microphone permission that works reliably

Signed releases now open the macOS microphone permission prompt correctly, with a clear way forward if the system cannot finish the request.

• The microphone permission prompt now appears when Sato Local needs it.
• A recovery message and Retry action replace the stuck state if macOS cannot complete the request.

Sato Local

Microphone access at the right moment

Sato Local now asks for microphone access as soon as your selected voice model is ready, before your first dictation.

• Voice setup clearly shows when microphone access is needed.
• Denied or restricted access includes a direct path to System Settings.

Sato Local

Voice setup that stays out of your way

Sato Local is easier to set up and quicker to return to. Model preparation now continues in the background with clear progress and recovery controls.

• See each preparation stage, elapsed time, and what Sato is doing next.
• Cancel, retry, or recover from a timed-out setup without restarting the app.
• Start or stop voice input from an active prompt with Command-Shift-M.
• Completed optimization is remembered, which makes later launches faster.

Sato Local

Private voice input, processed on your Mac

You can now dictate questions with WhisperKit and keep the entire recording and transcription flow on-device.

• Choose Fast for lower latency or Accurate for the best recognition quality.
• Download only the speech model you choose, then use it offline.
• Use your active Context Profile to improve recognition of technical vocabulary.
• Edit every transcript before anything is sent to your AI provider.

Chat

Sidebar text that fits your setup

Chat text remains readable in macOS Light appearance, and you can choose a color that works with your desktop.

• Improved sidebar text contrast in Light appearance.
• A persistent color control lets you choose your preferred chat text color.

Conversations

Add a fresh screenshot to any follow-up

An open conversation can now follow what is changing on your screen, not only the screenshot that started the thread.

• Press Ctrl+Option during an open chat to capture a new region for the next reply.
• Preview the pending image before sending your follow-up.
• See follow-up screenshots in chat history while earlier images stay in context.

Answers

Cleaner structure for longer answers

Completed responses now use explicit spacing for headings, paragraphs, lists, quotes, and code blocks.

• Long explanations are easier to scan in the chat sidebar.
• Math and chemistry notation remain readable after streaming finishes.

Answers

Markdown, math, and chemistry look right

Sato now renders common response formatting instead of showing the raw symbols behind it.

• Headings, lists, code, and emphasis render consistently.
• Common chemistry and mathematics formulas stay legible.

Conversations

Chat that moves with your workspace

Conversations can detach into a compact floating panel, dock to either display edge, and return from your recent history.

• Move and resize a detached chat without covering the material you are studying.
• Dock the panel to the left or right edge of any connected display.
• Reopen your five most recent conversations and pin the ones you need to keep.
• Choose GPT-5.6 Luna or GPT-5.6 Terra when using OpenAI.

Maintenance

Update prompts now stop when they should

A metadata correction prevents Sato 1.1 users from seeing an update prompt after they have already installed the latest version.

• No feature changes in this release.

AI providers

Choose the AI that fits the question

Sato expanded beyond Claude with support for OpenAI and Ollama, including local models that can keep screen data on your Mac.

• Switch between Anthropic, OpenAI, and Ollama from the menu bar.
• Set a different provider and model for individual Context Profiles.
• Responses gained their first pass of rendered Markdown support.
• Permissions and bundle identity became more reliable for future updates.

Good to know: If you upgraded from 1.0, macOS asked for Accessibility and Screen Recording permissions one more time because the bundle identifier changed.

Installation

Signed, notarized, and easier to open

Sato is signed with an Apple Developer ID and notarized by Apple, removing the unidentified developer warning from first launch.

• Clean first launch on macOS Sequoia and later.
• Existing users could install through Sato's built-in updater.

Launch

Sato's first public release

The original release introduced the screen-aware question flow and the small companions that make Sato feel at home on your desktop.

• Select any screen region with Ctrl+Option and ask a grounded question.
• Continue in a full chat with the original screenshot in context.
• Create Context Profiles for different subjects and workflows.
• Choose Max, Sky, Lexi, Rover, or Paris as your companion.
• Use Stealth Mode when you want the assistant without the sprite.