Prompt Assay
XI.Subscription

No demo call. No card. Free tier ships every AI instrument.

Self-serve through Team. Contact us only for SSO, BAA, or DPA. BYOK at every tier · including free.

Free
$0/ month

For trying the editor and seeing what the AI pair does to your prompts and skills.

  • 7 platform-funded calls to start (no key needed)
  • Personal workspace
  • 250 AI calls a month, enough to explore
  • Every in-editor AI action on prompts
  • Skills authoring + multi-provider Behavioral Eval
  • Evaluation suites with LLM-as-a-judge scoring
  • 7-day version view
  • Single seat
  • BYOK: bring an Anthropic, OpenAI, or Google key
Start free
No card. Upgrade any time.
Most popular
Solo
$49/ month

For the prompt engineer who is shipping prompts and skills.

  • Unlimited AI calls on your keys
  • Full prompt + skill version history
  • Every in-editor AI action
  • Skills authoring + multi-provider Behavioral Eval
  • Evaluation suites with regression on every version
  • Public REST API + TypeScript SDK (prompts + skills)
  • Single seat
Start solo
BYOK required. Cancel any time.
Team
$99/ seat / month

For teams reviewing prompts and skills like code.

  • Everything in Solo
  • Shared org library for prompts + skills
  • Owner, admin, member roles
  • Shared evaluation suites across the team
  • Shared Skills + Behavioral Eval reports across the team
  • Invitations and seat management
  • 3 to 15 seats
Start a team
Minimum 3 seats. Annual billing available.
Enterprise
Contact us

For engineering orgs that need SSO and custom controls.

  • Everything in Team
  • SAML SSO
  • Unlimited seats
  • Custom tier controls
  • Data processing agreement
  • Priority support
Talk to us
Custom contract. Talk to us.
BYOK at every tier. Your keys. Your bill. No markup.
Why this price

Some tools cost less. Here is what you get that they do not ship.

Six in-editor AI actions

Brainstorm, critique, improve, rewrite, compare, and convert. The full pair on prompts and on Agent Skills, not a single "AI suggest" button bolted on the side. Convert turns a plain prompt into a complete Agent Skill bundle in one shot.

Six-dimension critique with rationale

Clarity, completeness, structure, technique usage, robustness, efficiency for prompts. A parallel six-dimension scorecard for Skills (Discovery Fidelity, Instruction Quality, Example Coverage, Cross-Provider Portability, Token Efficiency, Security Posture). Written reasons, not a single number.

Multi-provider Behavioral Eval for Skills

Run trigger and non-trigger probes across Claude, GPT, and Gemini in parallel. Score activation, instruction adherence, and cross-provider consistency. No vendor ships a neutral grader for their own format; we do.

Multi-provider authoring

Anthropic, OpenAI, and Google all wired into the same editor. Switch providers on a prompt or a Skill and re-run without changing tools.

BYOK as a principle

Your provider account, your bill, no markup on a single token. Most tools resell inference or proxy your traffic. We do neither.

Reusable context blocks

System instructions, examples, guardrails, task frames, all reusable across prompts. Edit once, propagate everywhere.

Self-serve, no demo gate

Free, Solo, and Team are all self-serve. Talk to us only for Enterprise controls.

If a cheaper tool covers what your team does, use it. We do not chase price.

XII.Marginalia

Questions, answered plainly.

Yes for ongoing use, on every tier. New accounts get 7 platform-funded calls (Critique, Improve, Rewrite) to explore the workbench. Connect a Claude, OpenAI, or Google key when you're ready to keep going. Inference is billed by your provider, never by us.

Open the workbench. Ship prompts and Skills that hold up.

Free to start. Your keys, your bill, no demo call.

Open the workbenchRead the docs →