Scout Persona
A near-free developer tool: upload a persona charter, paste a model API key, run it against our probe library, see exactly where your persona holds and where it drifts before you ship.
What you get
A standalone tool where you upload a persona charter (or paste a system prompt), provide a model API key, and run it against our probe library — invariant probes, adversarial probes, consistency probes. You get a structured report showing where your persona holds and where it drifts.
What makes it different
Most teams test their AI persona by talking to it and trusting their gut. Scout Persona is the discipline of testing it against probes designed to break it — at near-zero cost, in your own environment, before you ship. It's also a marketing surface for our deeper Synthetic Personality research work.
What it can do for you
Catch personality drift before it ships to users. Stress-test your brand voice against adversarial prompts and edge cases. Compare persona-charter versions side by side. Validate that a model swap doesn't change behavior. Give your team a discipline they didn't have.
- Free for individual developers; paid for team accounts
- Self-serve, web-based, no engagement required
- Invariant + adversarial + consistency probes
- Compare two charter versions side by side
- Model-swap testing (does behavior change across providers?)
- Structured report with deviation scores
- API for CI integration
- Funnel into Synthetic Personality service if needed
Who it's for
AI engineers, prompt engineers, and anyone shipping a persona-based AI who wants confidence that the persona holds. Solo developers and small teams who can't justify the full Synthetic Personality engagement but want the discipline.
How we ship it
Self-serve. Sign in, paste charter, paste API key, run. Account creation is the only barrier; the rest is automatic.
Frequently paired with
Related products
A one-week engagement that maps your AI-readiness — what's in your data substrate, what's missing, what AI capabilities you could deploy today, and what would need to be built first.
We take 60 days of your production AI traces, build a real eval set, and hand it back with a working harness — so you can finally measure whether your AI is getting better or quietly worse.