AI that's actually private
June brings dictation, meeting notes, and a local agent into one workspace. Your data stays on your Mac, and AI requests route through a backend you can verify.
Stop uploading your life to AI companies
Every prompt, file, and workflow reveals sensitive data about you. Most AI apps ask you to trust them with all of it.
AI for everyday life and work
June lets you use AI without compromising on privacy
Chat with AI
Ask questions, do research, brainstorm ideas, build a plan, while keeping it all private.
On it — going through your Downloads folder. It all stays on your Mac.
Found 14 receipts from May totaling $2,847.10. Here's the breakdown:
- Groceries — $612.40
- Dining & takeout — $488.20
- Subscriptions — $214.97
- Transport — $176.50
- Shopping — $1,355.03
Want me to export this, or dig into any category?
Speak instead of type
Press the keyboard shortcut and talk. June turns your voice into polished writing, in any app on your computer.
Hi Maya,
Thanks for sending the revised timeline. I went through the milestones this morning, and Friday at 2pm still works well for me.
I’m aligned with using the review to settle scope, risks, and the handoff plan before the prototype review next week.
Sam
Effortlessly capture meetings
June can detect supported meetings and offer to write the notes for you, without a bot ever joining the call.
Q3 Roadmap review
The team aligned on shipping the agent beta to design partners next week, and holding pricing until the security review lands.
Decisions
- Beta to design partners on Thursday
- Hold pricing until the security review lands
Action items
- Book the security review
- Send the design-partner list
- Move launch checklist into the roadmap folder
Risks
- Security review may slip by one day
- Waitlist copy still needs legal approval
Context
- Design partners want earlier access to the import flow
- Support team needs a concise beta FAQ before invitations go out
Let June do the work
Schedule tasks, create reports, and prepare research. June simplifies your workflows without training someone else's model.
I found 18 upcoming renewals, scored them against ticket volume and product usage, then drafted the weekly report with account-level next steps.
Highest risk this week
Northstar Labs - $52K ARRMeadow Health - $41K ARR
weekly-risk.md
Weekly renewal-risk report
Generated every Monday at 8:00 AM
4 accounts moved into elevated risk this week
Highest risk
Usage down 38% · 6 open tickets
Champion left · renews in 21 days
Import failures · 3 seats
Recommended follow-up
- Schedule the Northstar recovery call
- Send Meadow the Q3 admin controls preview
- Attach Vale import logs to eng triage
How June keeps it all private
Architecture built to be private by default.
Local by default
June's agent runs locally on your desktop, built on the open-source Hermes framework. Files, sessions, memory, and state stay local by default.
Using private models
June defaults to private Venice models with zero data retention: nothing stored, no training. Anonymized third-party models are opt-in.
Minimal data retention
June stores only what it takes to run the service: account, login, and billing records. Your prompts, transcripts, and memory are not on that list.
Open-source and verifiable
June's code is open source and the backend runs in a TEE, so you can verify what runs on our servers and confirm we don't store your content or train on it.
AI with your privacy in mind
Claims most AI apps can't make, and an architecture that lets you check them.