Get to know MO.
A personal AI companion taking shape around conversation, useful memory and a charm you can make your own.
A little memory. A little help. A companion of your own.
A little context goes a long way
Carry the conversation forward.
MO’s memory is being designed to keep selected useful details after your initial permission. You would be able to review, correct and delete them.
Help at a useful moment
A reason to bring it up.
We’re exploring how MO could connect details across conversations and offer suggestions when they’re relevant. Their usefulness still needs testing.
Your question, your camera
Ask about what you see.
The planned camera experience starts with you. Ask about an image when you choose; phone or cloud processing may be used with permission.
Personal memory
Keep what helps. Change what needs changing.
Selected details, with permission
The plan is to remember useful details selectively after your initial permission, with controls to inspect each memory.
Make a correction
Your explicit corrections would update the current memory. Other unresolved conflicts would remain marked as uncertain.
Know whose detail it is
MO is being designed to distinguish who said something from who it concerns. Unclear attribution would stay unassigned.
Keep secrets out of memory
Passwords, verification codes and sensitive personal details are excluded from the planned persistent memory. Filtering still needs testing.
Raw audio is temporary
The design calls for raw audio to be deleted after processing. Selected memory details would remain until changed or deleted.
A companion with character
A little presence in your day.
We’re exploring a portable companion with expressions, short written replies and customizable charms and accessories.
Your controls
Decide what MO keeps.
Review and delete
Planned controls would let you inspect remembered information and remove what you no longer want stored.
Pause memory
Turning memory off is intended to stop new records and the use of existing memories. Stored records remain until you choose to delete them.
Choose when to ask
Image questions start with you. Optional cloud image permission would be revocable, and continuous capture is outside the plan.
Local core
Designed to work close to you.
Core processing on MO
Speech recognition, generated text and personal memory are required to run on MO itself. Hardware and model choices remain open.
English and Turkish
Both languages are part of the product plan and the planned initial studies. Those studies have not started.
Replies on screen
MO’s screen is intended to show expressions and short meaningful text. The first version does not include spoken replies.
A device still taking shape
Battery life, response time, heat and manufacturing cost still need physical testing. Pricing and availability haven’t been announced.