Pricing
Decide more. Build less. One unbuilt weekend pays for a year.
Every tier ships the rubric and the deterministic verdict — that part is table stakes. Paid tiers unlock the live evidence pipeline, founder-specific calibration, and the MCP endpoint your build agent can call. That's where Trigvale stops being a smarter prompt.
Live evidence
Six public sources (Reddit, GitHub, Hacker News, Stack Overflow, dev.to, Product Hunt) fetched per archetype, refreshed every 6h, with sparkline trends across every founder evaluating in the same archetype.
Founder calibration
Your skill + distribution graph re-weights every score. Verified GitHub stats and 30-day public commit cadence outrank declared signal — private repo data when GitHub is linked via OAuth.
Archetype context
Each idea joins 1–3 archetypes from a 16-cluster catalog. Corpus-derived stats plus semantic neighbours (embedding cosine within the cluster).
Agent endpoint
Callable from Cursor, Claude Code, and any MCP-aware build agent so validation can’t be skipped.
Explore
Try the rubric on three ideas. No card required.
3 evaluations, lifetime grant
- 3 evaluations — one-time, no monthly reset
- 10-dimension scorecard + deterministic verdict
- Idea archetype assignment
- No live evidence, no founder calibration, no sprints
Organize
Idea vault + verified skill graph.
About $0.50 per scored idea
- 50 evaluations / month
- Source-tagged evidence brief (observed / inferred / AI / claim)
- Founder skill graph with auto-verified GitHub signal (repos, stars, 30-day commit cadence, languages)
- Per-cluster archetype context with corpus-derived stats + recent neighbour ideas
- 10 validation sprints / month
- Markdown export
Validate
Live evidence + sparkline trends.
About $0.40 per scored idea
- 200 evaluations / month
- Live market evidence from 6 sources (Reddit, GitHub, Hacker News, Stack Overflow, dev.to, Product Hunt), refreshed every 6h
- Per-archetype sparkline trends across all founders evaluating
- Unlimited sharpening within an idea
- 50 sprints / month
- Decision memo export
Integrate
MCP endpoint for your build agent.
Unlimited evaluations + agent API
- Unlimited everything in Validate
- MCP server for Cursor / Claude Code / Codex / ChatGPT
- Agent API tokens callable from any IDE or automation
- Webhooks + structured JSON exports
- Priority support
Subscriptions are billed monthly via Stripe. Cancel from the customer portal at any time. The rubric and the deterministic verdict are identical across plans — paid tiers add the data layer (live evidence, founder-specific calibration, archetype context) and the agent endpoint. One weekend of unbuilt wrong-idea work pays for a year of Validate.
Credit packs
No subscription. Pay once, score when you need to.
Got one idea in mind, not twenty? Buy a pack of evaluations, use them within 12 months. Stacks on top of your free or paid quota — credits cover any overflow.
Starter pack
15 evaluations, one-time payment
$0.60 per scored idea
- 15 evaluations
- Use within 12 months
- Stacks on top of free / paid quota
Standard pack
40 evaluations, one-time payment
$0.48 per scored idea
- 40 evaluations
- Use within 12 months
- Stacks on top of free / paid quota
Power pack
100 evaluations, one-time payment
$0.39 per scored idea
- 100 evaluations
- Use within 12 months
- Stacks on top of free / paid quota