Every plan is the full product.
Weekly agent evaluations, live developer tracking, and recommendations you can act on. Plans differ on two things only: how many audiences you evaluate and how much traffic you track.
Weekly agent evaluations for every ICP, plus human evaluations on demand. Find out what breaks before your users do.
The tracking script. Time to value, journey stages, friction signals, and session timelines from real developers in your product.
Recommendations backed by evidence from both layers. Each one comes with a deliverable type and a quote, so you can buy the ones you want written and shipped for you. Available as an add-on on any plan.
You pay for audiences and traffic. Nothing is held back.
One product, one core audience. Everything the platform does, focused on the developers who matter most right now.
- ✓ 1 ICP with a weekly agent evaluation
- ✓ 1k developer sessions per month, each reconstructed into a complete journey with an AI summary
- ✓ Full platform: live tracking, adoption score, recommendations
- ✓ Human evaluations at $499 each
Your product serves more than one kind of developer. Evaluate five audiences separately and see which ones your product is actually ready for.
- ✓ 5 ICPs, each with a weekly agent evaluation
- ✓ 10k developer sessions per month, each reconstructed into a complete journey with an AI summary
- ✓ Full platform: live tracking, adoption score, recommendations
- ✓ Human evaluations at $499 each
Broad surface area and serious traffic across many audiences, with room to evaluate all of them every week.
- ✓ 15 ICPs, each with a weekly agent evaluation
- ✓ 50k developer sessions per month, each reconstructed into a complete journey with an AI summary
- ✓ Full platform: live tracking, adoption score, recommendations
Included in every plan
Every plan is the complete product. You scale by how many audiences you evaluate and how much traffic you analyze, never by unlocking features.
Real developers, on any plan.
Agents catch what breaks. Humans explain what it feels like. We match developers from our network to your ICP, they record their screen trying your product for the first time, and they say exactly what they think. You get the recording and a findings report.
Buy the fix, one recommendation at a time.
A result is one recommendation, implemented for you. Every recommendation comes with a deliverable type and a quote, so you buy only the ones you want shipped. Results are an add-on on any plan, never bundled.
We write and ship the words. We do not modify your application.
Publishable content we write and ship as a pull request to your site or docs repo: documentation, blog posts, messaging, pricing copy, email copy. Each one is quoted per recommendation. Most results run $499 to $1,499.
For product or application work we do not implement, a document specifying exactly what to change, why, and in what order, grounded in the evidence behind the recommendation. Flat $999.
Results are available now as an add-on on any plan. Buy the recommendations you want shipped.
Common questions
What's an ICP?
An ideal customer profile: a specific developer audience defined by stack, role, and experience. A backend engineer at a startup evaluating your API is a different ICP than a data engineer at an enterprise. Each ICP on your plan gets its own agent persona and its own weekly evaluation. During onboarding we detect your likely ICPs from your product and you confirm or correct them.
What's an agent evaluation?
Every week, an AI agent built to match one of your ICPs tries your product end to end: it reads your docs, signs up, follows your quickstart, and works toward first value the way a developer's coding agent would. Each run produces a scored report showing where it got stuck, what it misread, and how long value took. Reports arrive by email and live in your dashboard. Agent sessions are labeled in your data and excluded from your real-developer metrics.
What happens if I go over my session cap?
Tracking never stops. Every session keeps recording and raw timelines stay available, no matter your volume. Above your plan's cap, AI session summaries pause until your next billing cycle or an upgrade. You lose narration, not data.
How do human evaluations work?
$499 each, on any plan, never bundled into subscriptions. We match real developers from our network to your ICP. Each one records their screen trying your product for the first time, narrating what they expect, what confuses them, and what they'd do next. You get the recordings plus a findings report that synthesizes the patterns across sessions. A 3-pack is $1,299.
What is a result?
A result is one recommendation, implemented for you. Every recommendation comes with a deliverable type and, when it's serviceable, a quote. Buy the ones you want shipped. Results are an add-on on any plan, drawn from your agent evaluations, human findings, and live signal, and available now.
What's the difference between a content pr and a strategy doc?
A content pr is publishable content we write and ship as a pull request to your site or docs repo: documentation, blog posts, messaging, pricing copy, email copy. A strategy doc is for product or application work we do not implement. It specifies exactly what to change, why, and in what order. We write and ship the words. We do not modify your application. Some recommendations are advisory only, guidance for your team with no purchasable deliverable.
How is a result priced?
Quoting is automatic. The recommendations engine sizes each serviceable recommendation and quotes it. A content pr is priced per recommendation, and most results run $499 to $1,499. A strategy doc is a flat $999. Advisory recommendations carry no quote.
How is a result delivered?
A content pr ships as a pull request to your repo. A strategy doc is emailed to you and downloadable in your dashboard. Every result is emailed and in the dashboard, delivered within 3 to 5 business days.
Your first agent evaluation runs during onboarding.
Pick a plan, confirm your ICPs, and see what an agent finds in your product this week.