PRICINGFree answers the launch question. Paid buys volume.
Almost nobody runs 50 apps. Free watches one, includes graphs and the GitHub launch gate, and allows 500 SDK events per project per day. Pro and Business raise the caps and unlock your own AI key. They do not unlock a different verdict.
PLANFree
Free
- Sign in, then paste a public https URL you own. Ready / Fix first / Do not launch from stored evidence
- 10 public-app checks per month (24-hour cache, no extra AI spend)
- Watch 1 live app once a day, with a weekly email
- 1 connected GitHub app. 2 seats so a cofounder can join
- Quality Hub: cases, suites, API checks, requirements. Unchecked is not Ready
- Playwright, Cypress, Selenium, or Robot on GitHub-hosted runners that Invince starts. Not a Cloud Run browser grid
- 4 exportable software documents per app (tables from the repo; AI pages optional)
- 500 SDK events per project per day
- 8 shared AI requests per month. Improve ranking does not spend them
- Graphs, launch gate, Passport, and Improve ranking (same product as Pro, smaller caps)
Start freeRECOMMENDEDPro
$19/mo
- 50,000 SDK events per project per day
- 300 shared AI requests per month, plus your own AI key
- Agent Gateway (MCP) so your coding agent can read findings
- Watch about every hour on the apps you ship, with change emails and a weekly digest
- Same launch gate, Quality Hub, and graphs as Free, on more connected apps
- Playwright / Cypress / Selenium / Robot on GitHub runners plus owner k6 and lcov. Load still does not fire from Invince servers
- Control-loophole review, claimed-fix rechecks, and docs that follow the repo
- A reply when runtime errors pop up (SDK on your app)
- Agent cost, tokens, and stuck-loop graphs from the Node SDK
- 40 public checks per month, plus alerts before TLS expiry
- 3 connected apps and 5 seats. enough for a product, not a museum
- 12 software documents per app, including handbook pages grounded in the repo
- Your own AI key when the shared quota runs out
Upgrade to Pro · $19/moPLANBusiness
$49/mo
- Same depth as Pro on every connected app, including Quality Hub and owner-run browser JSON ingest
- Watch every 15 minutes (scheduler floor, not 5 seconds or PagerDuty)
- 10 Watch slots and 10 GitHub apps if you truly run that many
- 12 seats and 100 public checks per month
- 200,000 SDK events per project per day
- Still not 50 empty dashboards. Extra slots are for a small portfolio.
- All 20 software documents per app
Upgrade to Business · $49/mo ONE APP, MORE TO GIVEConnecting one app unlocks the product.
Health over time
A scored graph with axes after each completed scan. not a 2-pixel sparkline that hides until day two.
Findings vs issues
Open security findings and runtime issues on the same scale so you can see which way the app is moving.
Hits, logins, bots
First-party SDK traffic: who showed up, who signed in, and whether something non-human is hammering you.
What people do
Named actions as a real bar graph. Clicks and captureAction() fill it.
Launch gate
Ready / Fix first / Do not launch on the pull request, from the same evidence as the dashboard.
Watch + error replies
Scheduled public-page graph and email (daily Free, hourly Pro, 15 minutes Business). When an error pops up, a suggested reply is waiting. without pasting the stack to the user.
Agent spend and loops
startAgentTrace() on the Node, browser, and Python SDKs. Seven-day cost and token charts, loop flags, and a notice when a run retries the same step.
What is actually built.
The loops exist in the product. They are ready to charge for once the GitHub App and the browser snippet are on the app you own. We do not sell visibility into other people’s logins.
Watch (public app)
Ready to sell
Scheduled re-check (daily Free, hourly Pro, 15 minutes Business), change email when the verdict or a functional page break moves, weekly quiet note, and a response-time / findings / verdict graph. Hashed CSS class churn is ignored. It sees the public page Invince itself fetches. not other people’s sessions.
Launch gate
Ready when GitHub Checks are on
Writes a check run on pull requests. Needs the GitHub App with Checks: write and pull_request events. Require that check in branch protection. An unfinished scan stays in progress, not a green neutral. The Launch page still records the verdict if GitHub rejects the check post.
Loopholes, claimed fix, docs drift
Ready to sell, with honesty
Loopholes are conservative review prompts with file evidence, not a proof of exploit. Claimed fixes stay open until the next scan confirms. Docs flag when a commit moves the facts.
Traffic, logins, bots, error replies
Ready with the SDK on your app
Built end-to-end: ingest, hourly rollups, bot watch, suggested reply. Not for sale as “see any website’s logins.” Public Watch cannot do that; the snippet on the app you own can.
Agent observability (Datadog for your agents)
Ready with startAgentTrace()
Cost, tokens, and stuck-loop graphs for the agents you run. Invince does not execute the model. You report the run; we chart spend and flag the same step repeating.
Pro is $19/mo. Business is $49/mo.