Smart routing
Match task complexity, ambiguity, scope, and reasoning depth to a capability tier.
AI-powered model routingThe right model.
The right model.
For every task.
Rung breaks your PRD into editable tasks and asks your configured AI provider to recommend the right reasoning tier for each one.
Your PRD is parsed locally. Reviewed task text is sent only when you start live analysis.
PRD to tasks AI analysis Human review Export plan
PRD.md
Parsed locallyLightBounded changes
StandardCoordinated work
FrontierDeeper reasoning
Modeled costUnavailable · pricing not configured
Why Rung?
Why Rung?
For developers, students, and solo builders who want useful AI without opaque routing decisions.
Cost clarity
See modeled costs only when prices and token assumptions are configured.
Full control
Review, edit, or override a recommendation before you hand it off.
Export ready
Get a clear Markdown routing plan with rationale and review flags.
A simple four-step flow
From PRD to routing plan.
Review every task. Adjust every recommendation. Export when you’re ready.
Add your product requirements.
Edit extracted tasks before analysis.
Inspect and override suggestions.
Download routes and rationale.
Powered by hosted AI models
Choose the right capability tier.
Rung uses the configured AI provider to analyze each reviewed task and recommend one of three capability tiers. Recommendations are validated and remain under your control.
Light
A capability recommendation for smaller, contained tasks.
Standard
A recommendation for tasks that span connected steps.
Frontier
A recommendation for tasks needing more reasoning depth.
Build with intentionBuild smarter.
Build smarter.
Route with clarity.
Start from the requirements you already have. Keep every AI-assisted decision clear and reviewable.