Meet June: your private AI agent now with local models on your Mac.
Turn the wifi off: Agent, dictation, meeting notes, and your files all in one workspace.
100% free to use with local inference.
Open Source. MIT. Local first.
Try it: https://t.co/SC6KoHApdF
@omniscient57@OpenSoftwareCo our inference for open-weight models are run by Venice's private infrastructure, not the model creators themselves.
we detecting meetings by checkingย if you're in a browser/Teams/Zoom/etc and your Mic is being used. we'll teach this info to June for next time โย thank you!
AI Meeting note companies have been stealing your private meetings for a long time to train new models (look up the lawsuits). This is not only illegal, but a major invasion of your privacy and an insult to your humanhood.
Your private meetings should stay private.
We've just shipped private, automated meeting notes on your Mac, without a bot on the call.
June detects meetings and takes notes for you.
Turns it into decisions, action items, and follow ups.
It all stays private. Stored locally.
Open Source. MIT.
Try it for free.
Introducing private voice to text dictation on your Mac.
Hold the shortcut, talk, and June types the clean version of what you said into any app.
Routed through @askvenice zero-retention models by default.
Your thoughts stay yours.
Open Source. MIT License.
Read more:
We've all heard of horror stories of desktop agents bricking computers, unnecessarily modifying workspaces, and doing things it shouldn't be doing.
June's sandbox mode allows read capabilities while limiting write actions to its own sandbox to let you harness the power of AI agents safely.
Private AI, sandboxed locally on your Mac.
June's agent now starts in Sandboxed Mode by default.
It gives your agent a protected place to work in.
June can use the context it needs, but it canโt freely write across your mac.
Open source. MIT.
How it works โ
Your June agent keeps files and memory local while running inference in Venice's cryptographically attested private cloud because the Hermes framework executes on your Mac but routes LLM calls to zero-retention models like PrivateGLM 5.2 in TEEs, with anonymized third-party models available as opt-in. The compute is powered by DIEM (Venice's tokenized API credit at 1 DIEM = $1/day), though you access this via OpenSoftware's Hobby/Pro subscription at $20/month rather than holding DIEM directly. So you get local data sovereignty with verifiably private, token-backed inference.
it's crazy to see sign ups roll in for the product but not be able to see anything they're doing because it's actually private. if you run into a bug can you please let us know ๐
Meet June, private AI on your Mac.
June bundles it all in one:
- An agent + chat
- Voice dictation into any app
- Automated meeting notes
- Private OSS & anonymized frontier models
Files, context, and data stay yours.
Open source. MIT.
Try it: https://t.co/pDrMMBVpSd
June couldn't have been built without Venice.
Open source models are extremely powerful, but consumer hardware can't run them. We needed a cloud provider that committed 100% to privacy, and Venice was the clear choice.