Focused instructions.
Each AI step gets your context and a defined job.
Your context in.
A clearer day out.
What you share
- You writeLaunch the workshop FridayYou writeI need lighter evenings
- From Gmail“Is the $49 checkout ready?”From Gmail“Could you send the booking link?”
- Repeating taskEvery Friday: review bookingsRepeating taskWeekly: update attendees
- You add a taskDraft the launch emailYou add a taskCreate a promo graphic
- From Google Calendar17:30 Client call endsFrom Google Calendar45 min Free this evening
- Pasted from ClaudePast attendees asked for morePasted from ClaudeLead with a hands-on exercise
Context pulled from your connected tools
+29 more
What you get back
- Next moveTest checkout20 min
- Insight85 min of work. 45 min available.
- Helpful ruleLonger work: a free morning
01 / How it worksFollow one note
Follow one note
through the Lab.
Scroll. Each step is one thing the Lab does for you, on the same example.
- Step 1
You write it down
One messy note, in your own words. No forms, no fields.
- Step 2
It reads
It finds your goal, your deadline and the time you have.
- Step 3
It sorts
The note becomes clear tasks, each with a type and a time.
- Step 4
It picks
It decides what comes first, and tells you why.
- Step 5
One next move
Start here, with the first steps spelled out. The rest wait in line, in order.
- Step 6
It follows through
Finish a task and it spots what comes next: a promise to keep, or a smaller first step for something that keeps slipping. You say yes or no.
- Step 7
It remembers
Your note, a summary pasted from Claude or ChatGPT, a quick correction. Tell it once and it sticks.
- Step 8
It learns your rhythm
It spots a pattern and suggests a rule. Say yes, and tonight's plan already follows it.
An example, not a live AI run.
02 / UNDER THE HOOD
AI with checks built in.
Checked responses.
Code checks the response format. Your plan changes only when you request a re-plan.
Tested workflows.
Evaluation cases check planning behavior. Run logs track the model and response time.
03 / YOUR CONTROL
You stay in charge.
The Lab handles
- ✓Sort tasks and estimate time
- ✓Suggest what to do next
- ✓Fit tasks to your available time
- ✓Suggest follow-ups and rules (preview)
You decide
- ↗Choose a different next move
- ↗Accept or dismiss a suggestion
- ↗Correct an estimate or task
- ↗Set the rules in your own words
Wrong suggestion?
Check the reason, choose another task or correct the estimate. The Lab can miss context; you can change the plan.
AI providers process relevant inputs. Export or delete your data in Settings. Encrypted backups may remain for seven days.
Your data, explained ↗Questions?
One assistant, or a team of agents?
You work with one assistant. Behind it, the Lab coordinates AI reasoning, planning and code checks. Different steps can use different prompts or models.
Which AI does the Lab use?
Models run through OpenRouter, with a primary model and an optional fallback. We choose the models and write the prompts.
Does it read my email or other AI chats?
No. You choose what to share. Pasting an email or AI summary does not connect the account. Direct integrations are coming.
Is my data used to train AI models?
AI routing excludes providers that collect data for training. Relevant inputs go to providers to process your request. See our privacy policy for storage and retention.