What Yomi is
Yomi is an AI assistant that runs in your Windows system tray. It can see your screen, hear your voice, and act across your apps, so you can ask questions about your work in plain language instead of switching windows and copy-pasting.
You reach Yomi two ways: the desktop overlay on your PC, and the Telegram bot when you're away from it. Both share the same memory and connected apps.
Getting started
- 1Download for Windows. Grab the installer and launch Yomi. It lives in your system tray.
- 2Sign in. Continue with Google or GitHub. No password to manage.
- 3Connect your apps. Link Gmail, Calendar, Drive and others from the dashboard with one click each.
- 4Link Telegram (optional). Use the dashboard's secure link flow to chat with Yomi from your phone.
macOS is planned; Windows is available today.
The desktop assistant
Yomi sits quietly in your tray and appears as a floating overlay when you call it. It reads the screen you point it at, so answers are grounded in what you're actually looking at.
Screen-aware answers
Ask about whatever's on screen, like a doc, an error, or a dashboard, and get a grounded answer.
Voice & text hotkeys
Trigger Yomi by hotkey, then type or speak. A fast path replies in about two seconds.
Visible capture states
You always see when Yomi is capturing the screen, so nothing happens silently.
Two routing paths
Simple questions take the quick path; anything needing your apps runs the full agent.
Telegram bot
Connect Telegram from the dashboard's secure link flow, then message Yomi like any chat. It handles text, voice notes (transcribed automatically), and images/screenshots for analysis.
| /new | Start a fresh conversation (clears the current chat context) |
| /stop | Stop whatever Yomi is currently doing |
| /pending | Show actions waiting for your approval |
| /approve | Approve the pending action (or pick one if several) |
| /deny | Reject the pending action |
| /help | List the available commands |
App connectors
Connect the tools you already use. Yomi requests OAuth access (or an API key) and only acts when you ask. Connect and manage them from your dashboard.
Memory & knowledge
Long-term memory
Yomi remembers durable facts about you and your projects, so it doesn't ask the same thing twice. Starting a new chat clears the conversation, never your memory.
Your documents (RAG)
Synced documents become searchable context, so answers can draw on your own material.
Voice & vision
Speak and listen
Voice notes are transcribed, and Yomi can reply with a spoken voice message when you ask.
Image & screen analysis
Send a screenshot or photo and Yomi describes, reads, or reasons about what's in it.
Approvals & safety
Actions that send or change things, like sending an email, pause for your approval first. Review the preview, then approve or deny (on Telegram, use /approve and /deny). Yomi never sends on your behalf without a confirmation.
Plans & credits
Every interaction draws from a single credit balance. Each plan includes a monthly credit allowance, and your dashboard shows remaining credits, recent activity, and the next reset date. When you run out, Explore upgrades to a paid plan and Pro/Max can top up with credit packs.
25 credits
30-day trial to try everything.
2,500 credits
Higher limits + credit packs.
10,000 credits
Highest limits for heavy use.
Credits abstract away the underlying model, voice, vision, memory, and connector costs so Yomi can improve routing without changing the dashboard experience.
Privacy
- Screen capture is always visible. Yomi never reads your screen silently.
- Connected-app data is used to answer your request, not stored beyond what's needed.
- Google Drive file contents are not retained.
- You can disconnect any app at any time from the dashboard.
Ready to try it? Create your account or download for Windows.