Turn more trials
into customers.

Sunboard helps your own AI guide the activation work: clarify the path to value, implement native checklists and tours, and connect the experience to the events that show users are ready to pay.

  • Build around the moment users reach value
  • Plan it with your own AI, like an expert guide in the room
  • Preview, launch, and improve flows against activation signals
Get started free
sunboard — zsh
Works with any harness
AI skills workflow

Four skills for activation work

Each Sunboard skill gives your AI a structured step: ask the right questions, wire the real product, preview the experience, and improve it against the signals that matter.

your AI — ~/acme-app
What Sunboard adds

Activation work your AI can reason through

Sunboard gives the AI you already use the product context, spec format, runtime, and launch controls to help more users reach value and become customers.

Start with the value moment

Instead of starting from a widget editor, your AI helps define the audience, activation event, and shortest useful path. The output is a structured spec you can review together.

Build guidance into the product

Checklists, tours, hotspots, and events are wired where the user actually works, so onboarding feels native and can measure whether the right action happened.

Preview and revise together

Walk the flow in your app before customers see it. Ask your AI to tighten the sequence, change copy, or adjust completion rules while the full implementation is still in view.

Launch, target, and improve

Sunboard keeps production controlled: review a candidate, choose who sees it, then watch activation and completion signals so the next pass is based on evidence.

Checklist UI

Use the default, or build with your UI

Start with Sunboard's built-in checklist, or run the checklist UI skill and let your AI rebuild it using your design system while Sunboard keeps the progress logic underneath.

Use the built-in checklist

<Sunboard.Checklist /> gives you a polished floating checklist as soon as Sunboard is installed. It reads your activation spec and handles progress, actions, and completion states out of the box.

Get started1 / 3
Create your account
Connect a data source
Invite your team

Or generate one from your UI library

/sunboard.checklist-ui has your AI inspect your components, icons, and styling patterns, then rebuild the checklist with your UI library. Same useChecklist() data contract, your product surface.

Inline cardSidebarNav dropdownEmpty state

Onboarding

Finish setup
2 / 4
  • Create your account
  • Connect a data source
  • Create your first boardStart
  • Set up alerts
Close the loop

See where users drop off, then fix the flow

Sunboard tracks the activation funnel behind each experience. Run /sunboard.analyze and your AI can reason from real drop-off data instead of vibes.

Measure the path to paid

Sunboard tracks views, step completions, and activation events for each experience and segment, so you can see who reaches the value moment.

Find the step that costs you customers

See the exact moment trial users stall, from first checklist view to goal event, instead of guessing from a pile of disconnected charts.

Turn diagnosis into the next experiment

/sunboard.analyze gives your AI the metrics and context to propose the next /sunboard.spec, /sunboard.tour, or /sunboard.implement change.

trial-activation
Last 30 days
Views1,284
Activation31%
Drop-off42%
Viewed checklist1,284
Connected a data source742
Invited a teammate430
Activated398

Biggest leak: “Connected a data source” → “Invited a teammate”. Team invite is blocking value — move it after the first board is live. /sunboard.spec

Ready to improve trial conversion?

Install the CLI once, run sunboard init, and your AI is ready to help more users reach value.

$npm i -g sunboard
$sunboard init

Free to start. Works with Codex, Claude Code, and opencode.

Questions founders ask

Have more questions? Reach out at sean@sunboardhq.com

Agent-native onboarding for SaaS. In-app checklists, tours, and hotspots that ship as code — authored by your AI coding agent, versioned with your commits.
Made in Sweden