Verification

Deployments & AI Verification

Deployed isn't done. DevSpec links every deployment back to the action items and conversations that shipped it — then AI browser testing opens the deployed app, walks the acceptance criteria, and records video proof that it actually works.

Test run — #214staging

Exponential backoff for Stripe webhook retries

AI browser test · video 0:48 · 11 screenshots

  • Trigger a webhook retry from the Stripe test dashboard
  • Retries arrive with exponential backoff and jitter
  • Event marked processed exactly once in the admin view
confidence 0.92pass — confirmed

— a test run, ten minutes after deploy. Nobody had to remember to check.

01The Problem

'Deployed' is where most tracking stops

The merge went green, the deploy went out — but did the feature work? On most teams someone checks manually, or nobody does. Bugs surface days later with no link back to the change that caused them, and 'done' quietly means 'probably fine'.

A feature deploys at 4pm. By 4:10, an AI test has walked through the acceptance criteria on the live URL and attached a video of it passing. The item flips to verified — nobody had to remember to check.

02The Mechanics

How it works

From webhook to verdict, the loop closes itself. AI browser testing is in early access — it's strongest on web apps deployed to a reachable URL.

Deployments link themselves

A webhook fires when your pipeline deploys. Because Autopilot and the plugin tag commits with [devspec:id], DevSpec connects the deployment to the exact action items it shipped — and marks them with a deployed badge.

Fig. i — Linked deploysapp.devspec.ai/changelog

The deploy knows what it shipped.

Every item arrives test-ready

Deployed items get a testability assessment: Ready, Needs Setup, or Not Suitable — with setup and cleanup scripts spelled out when the test needs data prepared. You can see at a glance what's verifiable.

Fig. ii — Testing briefsapp.devspec.ai/testing

Ready, needs setup, or not suitable — assessed before anyone asks.

AI tests in a real browser

An AI agent opens the deployed URL in a real browser and walks through the acceptance criteria the team wrote in conversation — recording video, capturing screenshots at each step, and scoring its confidence.

Fig. iii — Browser testapp.devspec.ai/testing

The same criteria the team agreed on, walked on the live app.

Verdicts close the loop

Confirm a pass and the item is verified. Confirm a failure and DevSpec spawns a bug item with the evidence attached. False positives and stale tests have their own paths — the record stays honest.

Fig. iv — Verdictsapp.devspec.ai/testing/history

A failure becomes a bug item with the video attached.

03An Index of Capabilities

Evidence, not optimism

01

Deployed badges

Items show exactly which deployment shipped them, on which target. The board reflects production, not intentions.

02

Full traceability

Conversation → action item → commit → deployment → verification. Every link in the chain is recorded and clickable.

03

Testability assessment

AI evaluates each item for automated testing up front, so you know what can be verified hands-free and what needs a human.

04

Video evidence

Each test run records the browser session. When something fails, you watch exactly what the agent saw instead of reproducing from a stack trace.

05

Failures become work

A confirmed failure spawns a bug item automatically — evidence attached, ready to queue for Autopilot.

06

Run history

Every test run is kept with its verdict, confidence, and artifacts. Reliability over time is a graph, not a feeling.

DevSpec

Convinced?

Set up your first project in under five minutes. Connect your repo, start a session, and see what your duck can do.

Start Your Project

14-day free trial · 1,500 credits included