Work in threads with any AI. Pick up any thread with nothing lost.
Start a thread on one model, continue it on another, hand it to a colleague next week. tabbr keeps the whole thread — every turn, every version, every decision — so the work never has to start over.
Start a thread — free Free for individuals, forever. No card.Draft a pricing proposal for Acme. Three tiers, annual discount, keep the enterprise tier under $50 a seat.
Here's a draft with Standard, Professional and Enterprise at $12 / $25 / $40 per seat, 15% off annual. I've noted where Acme's procurement terms may push back.
Stress-test the enterprise tier against Acme's 400-seat rollout. Where does it break?
At 400 seats the platform fee dominates. I'd move it into the per-seat price and add a volume band at 250+. Revised proposal attached as v2.
Acme's legal came back asking about the volume band. Why did we add it, and what did we consider instead?
The band was added in v2 after a 400-seat stress test on GPT-5 showed the platform fee dominating. The alternative considered was folding the fee into the seat price. Both are in the thread above; here is the reasoning in one paragraph for legal.
Three things to know
- Thread
- Where work is created, with intelligence. Any model. Switch models without losing the thread.
- Page
- What a thread produces. A permanent address and a full history, kept forever.
- Workspace
- Where your company keeps both — and where anyone with access can pick a thread up.
Who it's for
Individuals — your thinking stops dying when the tab closes.
Companies — your people's AI work stays yours: every model, one record, exportable any time.