Introduction
Reshot organizes accessibility work around durable evidence: capture a real experience, evaluate it, consolidate repeated findings into barriers, remediate and verify them, then update the affected conformance record.
You do not need repository access or a developer for first evidence. Start in Experiences with either:
- Hosted discovery for a public URL or sitemap
- Browser recorder for an authenticated or multi-step journey in your own browser
CLI, CI, and repository integrations are optional expansion paths after the initial evidence is representative and the team has calibrated its policy.
The working model
| Destination | Purpose |
|---|---|
| Command Center | Resume setup and act on stale, failing, or unowned work |
| Experiences | Discover, record, version, scope, and evaluate complete processes |
| Barriers | Consolidate occurrences, assign ownership, remediate, and verify |
| Components | Trace shared UI impact and component-level remediation |
| Conformance | Decide criterion status over explicit scope and evidence |
| Records | Retain and verify signed evidence records during the private pilot |
| Settings | Manage people, policy, integrations, data lifecycle, and billing |
First-evidence sequence
- Create or open a project.
- Open Experiences.
- Enter a public URL, or record an authenticated journey with the extension.
- Review the proposed checkpoints and complete-process scope.
- Run the initial evaluation across the representative context policy.
- Triage the resulting barriers; unresolved manual questions remain explicit.
Reshot reports ACT-style outcomes such as passed, failed, cantTell,
inapplicable, and untested. An unmeasured result is never presented as a
pass, and an automated run is not a conformance decision.
Add engineering automation later
After the baseline is stable, install the CLI to reproduce the same experience
in customer-controlled infrastructure and connect a repository when delivery
feedback is useful. Policy begins advisory. It may enforce only approved,
high-confidence new mechanical failures; human findings, cantTell, worker
failures, and usage reconciliation do not create a false accessibility failure.
Continue to Installation for all three capture paths, or follow Quick Start for the non-developer first-evidence workflow.