TodoXP is a tasks-and-habits app for people who already tried Todoist, Things, Notion, and a Notes app, and bounced off each of them. The tagline is "tasks that respect how humans actually work" — meaning the app doesn't punish you for missing a day, doesn't gamify-shame you, and doesn't try to be your second brain.
What it is
The minimum viable productivity surface for a founder who lives in their calendar:
- Tasks with habits-and-streaks built in, not as a separate sub-app.
- Lightweight accountability — one shared view, not a notifications firehose.
- Local-first so the app works offline and on a flight, syncs when there's a connection.
- Calendar-aware because pretending tasks don't compete with meetings is the failure mode of every other tool in this category.
Free of the genre's usual sins: no streak-shaming notifications, no "you're falling behind" emails, no XP system that makes Friday feel worse than Monday.
Why I built it
I tried every task app on the market for a year. None of them survived the second time my schedule got blown up by a client emergency. Either the app punished me for the lapse (streaks broken, dashboard turning red, "you're behind!" emails) or it absorbed the lapse so completely I lost the signal entirely (everything just rolls forward to tomorrow, forever).
TodoXP is the middle path: it shows you the lapse honestly, but it doesn't moralize about it. You missed Tuesday. Wednesday is fine. Move on.
How it ships
Built solo with Claude as a daily collaborator. Local-first architecture means the data lives in the user's browser by default; sync is opt-in via a small backend that doesn't see plaintext task content.
Stack: React PWA (installable, works offline), CRDT-based local-first sync, no native app store presence by design.
Why this matters for AI systems
The connective thread between TodoXP and the rest of the studio is the operator-shape principle. Most productivity tools are built for the optimistic case (you stick to your routine). The 3am-page case (you missed three days because a client emergency happened) gets ignored, and the tool starts hurting more than it helps.
The same instinct — design for the messy case, not the demo case — is exactly what AI ops surfaces need. AstroSense applies it to agent runs. PainSense applies it to clinical intake. TodoXP applies it to daily work. Same posture, different domains.