It expires in weeks, not years
A prompt that worked in January behaves differently in July. The model changed underneath it and nobody told you.
The verification layer for AI knowledge
Every AI workflow is silently expiring.
Doestep is the only place that proves what still works — on this model, right now — because Doesteppers ran it and logged the result.
Live verification
Extract structured data from a messy invoice · v3
Not yet verified on Sonnet 5awaiting reports
A new model shipped. The badge started unverified — then Doesteppers re-ran it and it went live. That flip is the whole product.
The problem
AI knowledge rots, and nobody tells you when.
Nothing is reproducible
“I built an agent that 10x’d my workflow.” Great — can anyone else run it and get the same result? Usually not.
Everyone claims, no one proves
The internet is full of AI skills nobody can verify. Claims are cheap. A logged reproduction is not.
How it works
One loop, and it never stops turning.
A playbook gets proven, a new model breaks it, the community proves it again. The model-release event is the heartbeat.
Read the loop in depth on how it works →
The badge, up close
A badge that speaks in facts.
This is something no tutorial site, GitHub repo, or Discord has: a status that’s computed from real reproductions and can’t be written by hand. Conflicting results are shown, never hidden.
- The result
- Worked, broke, or mixed — a status color and a glyph, never color alone.
- The model
- The exact model version it was run on. Freshness is relative to models, not the calendar.
- The freshness
- Recomputed on every report, on model release, and nightly. It decays on its own.
- The reports
- How many distinct Doesteppers proved it — the evidence behind the badge.
The states
Verifiedclaude-sonnet-5 · 3 days ago · 14 reportsNot yet verified on Sonnet 5claude-sonnet-5 · awaiting reportsMixed — see reportsclaude-sonnet-5 · 9 worked · 5 brokeBroke on step 4claude-sonnet-5 · last worked on opus-4.1AI draft · unverifiedno reports yet
Same object everywhere — on the site, in a README, in a tweet. Mono for the data, status color for the fact.
State of What Works
Sonnet 5 dropped. The badges started flipping.
When a new model ships, every affected playbook flips to “not yet verified.” We publish the only report built from real reproduction data — what survived, and what quietly broke.
the moment a model ships
Verifiedclaude-sonnet-5 · 3 days ago · 14 reports
No one edited anything. The evidence just aged out from under it.
- 214
- playbooks flagged
- 130
- re-verified so far
- 61%
- still work
- 51
- quietly broke
Who's a Doestepper
You don't just read guides. You prove them.
A Doestepper is anyone who runs the thing and logs the result. You don't join a space or a channel — you join people, by the tools you build with and the work you do.
Find your Doesteppers by tool
n8nZapierLangChainClaude CodeMakeCursor
…and by role
Sales-AILegal-AIContent-opsPrompt engineersData teamsFounders
412 Doesteppers verified this on Sonnet 5 — not 412 “members.”
The next model ships in weeks. Be the one who proves what survives.
Join the first Doesteppers. When the badges flip, you'll be the reason they flip back.
Join Doestep412 Doesteppers already waiting