Watch an AI agent think
No model — every run is a hand-written script you can scrub.
livefig. 01
First, it writes itself a plan.
data/loop.tsread-only
taskTests are failing in utils.test.ts — find out why and fix it.
mind1 plan
plan
- ·Run the tests
- ·Read the failing test
- ·Find the bug
- ·Fix it
- ·Verify green
action stream1/25 events
waiting…
context0%
0
/ 8,000 tokens
0.0 / 40s
pausedrun 01 / 03
space play · ← → scrub · 1–3 runs