NewPersonal Future Intelligence · powered by NVIDIA Nemotron
Tomorrow starts with you.
Explore the futures your choices could create. Talk with your future self, compare possible paths, and turn your ambitions into a plan you can act on.
TOMORA explores possibilities. It can't know your future, and nothing it shows is a prediction.
- NVIDIA NemotronOn Nebius Token Factory
- A private workspaceYour future self is yours alone
- Memory you can readPlain notes you can edit or delete
- Asks before actingNothing changes without your OK
One goal, several futures
See the roads before you take one.
Name where you want to go and TOMORA lays out clearly different ways there, each with its time, weekly hours, risk and the assumptions it rests on. Pick one to see its route.
Example
Work and study route
If you take this road: Take a paid role next to the field and study part time, moving step by step toward the goal to move into data analysis.
- 1.5–3 years
- ~8 h
- Lower risk
Holds only if
- You can find a paid role close to the field
- You keep 6–10 hours a week for study on top of work
How it works
From where you are to a plan you can act on.
TOMORA follows one journey. Each screen in the app serves a step of it. Pick a step, or watch it play.
Step 01 of 04
Understand me
Where are you now, and what's in the way?
Full-stack developer, two hours most evenings. No GPU, no publications yet.
Saved as notes you can read, edit or delete.
Explore different futures
Ask what if, and see what actually changes.
TOMORA uses AI to reason about which roads exist and what they depend on. The arithmetic (time, money, hours) is plain code, so the same inputs always give the same answer, and you can see when a number was recalculated rather than newly generated.
- What if I can only give five hours a week?
- What if I don't reach the first milestone?
- What if I have less money?
- What if I take a safer route instead?
- What if I study for another year first?
Self-study route
Recalculated1–2 years
Same as the road's own estimate.
Talk with your future self
A conversation with who you could become.
An AI persona speaking as a plausible future version of you, grounded in your goals and the road you chose. It isn't a real future you, and it says so when it's unsure.
Remembers what you allow
Only notes you haven't excluded are used, and every one can be opened, corrected or deleted.
Asks before it advises
Follow-up questions, trade-offs named plainly, and contradictions pointed out kindly.
Separates fact from guess
What you told it is marked apart from what it inferred.
Ends with a next step
Conversations turn into small plans, and nothing is added to your roadmap until you approve it.
Speak or type
Voice conversations work when a voice engine is set up on your server. Text always works.
I keep saying I'll start the course, then the week fills up.
You told me you have about ten hours a week, but your last three weeks went to your job. That's not a willpower problem; it's a calendar one.
Could two fixed evenings work, say Tuesday and Thursday? I don't know your shifts, so tell me if that's unrealistic.
Turn insight into action
Fewer, better next steps.
The road you choose becomes a short plan: the next three actions, this week's priorities, milestones with target windows and a review when they come due. Progress only moves when you complete a real step.
Next three actions
Small enough to start this week, tied to the road you chose.
Milestones and reviews
Target windows, not deadlines, with a check-in to ask what changed.
Plans you approve
Your future self proposes; nothing lands on your roadmap until you say yes.
Privacy and control
Private by design. What you share stays yours.
Memory you can read
What TOMORA remembers is a folder of plain notes. Open, correct, exclude or delete any of them.
Runs on your server
Your data lives in a local database and Markdown files. A cloud copy is off unless you turn it on.
Keys stay server-side
Model and voice keys never reach the browser, and saved keys are encrypted at rest.
Clear about the model
Requests go to NVIDIA Nemotron on Nebius Token Factory, or to a local model in private mode.
Voice only with consent
A voice twin needs a recorded statement of consent in your own voice, and can be deleted.
Nothing public by default
An access code can lock the whole app. Share previews show the brand, never your goals.
Questions
Frequently asked questions.
Does TOMORA predict my future?
No. It lays out possible roads and what each depends on. Every time and cost is a range, every road lists the assumptions it rests on, and none of it is a forecast of what will happen to you.
What is the future self?
An AI persona that speaks as a plausible future version of you, using the goals and context you've shared. It's a way to think things through, not a real person and not a prophecy.
Which AI does it use?
NVIDIA Nemotron, served by Nebius Token Factory, for reasoning and conversation. Without a key, a built-in stand-in model fills in clearly labelled templates. A private mode can use a model running on your own machine instead.
Where is my information stored?
On the server you run TOMORA on: a SQLite database and a folder of Markdown notes. An optional cloud copy in Supabase is off unless configured.
Can I correct or delete what it remembers?
Yes. Every memory is a note in the Memory vault. You can edit it, stop it from being used in conversations, or delete it.
Does voice always work?
Voice needs a voice engine key on your server, and voice cloning needs your recorded consent. When voice isn't available, everything works by text.
Can I try it without sharing anything?
Yes. Demo mode fills the app with sample data in your browser and never calls a model or the internet.