For developers going AI-native
Become AI-native.
With proof.
You ship with Claude Code, Cursor, Warp, and Windsurf every day — but you have no idea if you're actually getting better. Slait reads your real sessions, scores how you work with AI across 8 dimensions, and shows you exactly what to fix next.
Works with
Overall
77/100
Percentile
Top 14%
8
Scoring modules
~60s
From upload to scorecard
6
Supported tools
$0
To get your first score
The gap is widening
The devs pulling ahead aren't smarter.
They have a feedback loop.
Every prompt, retry, and dead-end is a signal — but it disappears the moment your session ends. Without a way to see your patterns, you keep repeating the same mistakes: vague specs, missed constraints, debugging by guess. A year goes by and you're still "pretty good at Cursor." Meanwhile, AI-native engineers are shipping in hours what used to take weeks.
How It Works
From session log to scorecard in 60 seconds
Three steps. No setup, no plugins, no waiting on a sales call.
Upload your session
Export a transcript from Claude Code, Cursor, Warp, Windsurf, Codex, or Copilot. Drop the .md, .txt, .json, or .jsonl file — that's it.
Get evidence-based scores
Slait reads every turn and scores 8 modules — Planning, Debugging, Tool Usage, and more — citing the exact lines from your transcript as evidence. No vibes, no guesswork.
Level up and compete
Track your scores over time and climb The Slate — a public leaderboard of devs proving they're AI-native. See exactly where you stand.
Go to The SlateWhy Slait
Three things you can't get anywhere else
See your blind spots
Find the habits costing you hours — vague prompts, missing constraints, the debug loops you can't feel from the inside. Every score is backed by the exact line in your transcript that triggered it.
Level up on purpose
Stop hoping you're getting better. Pick a module, ship a few sessions, and watch your scores move. The feedback loop senior AI-native engineers build for themselves — now built in.
Prove you're AI-native
Anyone can claim AI fluency. Your Slate profile shows real, scored sessions across the tools that matter. Share it with teams, recruiters, or your group chat.
Scoring Model
8 dimensions of AI-native engineering
Every session is scored across the skills that separate good AI users from great ones. Each module returns a score, evidence, and a concrete next step.
Planning
Scoping work into clear, AI-shaped tasks before you prompt
Debugging
Finding the real root cause instead of looping on symptoms
Constraints
Stating requirements, edge cases, and limits up front
Iteration
Refining prompts based on what the model actually returned
Correction
Catching and recovering when the AI goes off the rails
Tool Usage
Using Claude Code, Cursor, Warp, Windsurf to their full power
Repetition
Avoiding the same prompt patterns that already failed
Understanding
Reading AI output critically and knowing what to do next
FAQ
The short answers
Which tools work with Slait?
Claude Code, Cursor, Warp, Windsurf, Codex, and Copilot. If you can export a session transcript as .md, .txt, .json, or .jsonl, Slait can score it.
What does "AI-native" mean?
Engineers who treat the model as a real collaborator — scoping, prompting, and course-correcting with intent. It's a skill, not a tool. Slait measures it.
Is my code private?
Only sessions you upload are analyzed, and your scorecard is yours. Public ranking on The Slate is opt-in.
Do I need to pay?
No. Sign up, upload a session, and get your first scorecard for free.
Ready when you are
Find out what kind of AI engineer you actually are.
One session. Eight scores. A clear next step. Upload your transcript and get your first scorecard in under a minute.
For Teams
Hiring or upskilling for AI fluency?
Slait gives engineering teams an objective way to see how their developers — and candidates — actually use AI to ship. Leave your email and we'll show you how it works.
Feedback
Send us feedback
Tell us who you are, how to reach you, and what you think. We read every message.