Record the test, attach the CI run, send one link. Reviewers approve without an account.
The product tour could not load. Check your connection and try again.
Start in the browser. Sign in only when you share or sync.
Record a flow with the browser extension. You get a titled test case with steps, expected and actual results, and screenshots. Confirm it, or mark the issue.
Send Cypress, Playwright, Cucumber, or JUnit results. Tests that don't exist yet are created from the run, later runs keep them current, and failure screenshots, stack traces, and retries stay with the run they came from.
$ node scripts/dossier-cypress.mjs --output cypress-results.json
$ npx testdossier upload \ cypress-results.json ✓ failed screenshot · stack · retries ✓ optional source context frozen
Send one read-only link. Reviewers read the evidence, comment, and record a decision. No account, no seat.
Pay for the people who create and manage evidence, not the people who review it.
Free links record one reviewer's decision. Pro links take as many reviewers as the report needs.
Steps, screenshots, and the console and network activity behind the failure. Open the request that returned 500.
| # | Action | Expected | Actual | Result |
|---|---|---|---|---|
| 1 | step | expected | actual | PASS |
| 2 | step | expected | actual | PASS |
Evidence
|
||||
| 3 | step | expected | actual | FAIL |
|
|
||||
Step 3 · Check the order total
{
"error": "INTERNAL_SERVER_ERROR",
"message": "Cannot read properties of null (reading 'total')",
"customerEmail": "[REDACTED]",
"requestId": "req_01JY8QG3B7N8K2M6P4T0V9XZ"
}
Sensitive fields were redacted before saving.
Uncaught TypeError: cannot read properties of null (reading total)
Use TestDossier beside TestRail, Xray, Qase, or Jira when evidence handoff is the bottleneck. It replaces screenshot folders, Word reports, and approval threads—not your system of record.
Bug tools record failures. TestDossier shows the full test picture—including passes—so reviewers can make a release decision.
Explore supported imports and CI formats →No. Anyone with the link can read the report, leave comments, and record a verdict in the browser. Only people who author evidence need a seat. Free links last 30 days and take feedback from one reviewer; Pro links can be permanent and take unlimited reviewers.
Because a video is not a record. Nobody re-watches 14 minutes to check step 9 six months later. Here every step is a written action, expected result, and screenshot, so a reviewer scans it in a minute and an auditor can quote it.
The same case keeps a snapshot from every run, so history and pass rate accumulate instead of living in one page. And the link carries a verdict, a comment thread, and optional email-verified sign-off, and never breaks when the case is edited later.
Your existing test manager keeps the catalog, permissions, and governance. TestDossier packages the evidence for people outside QA: steps, screenshots, CI runs, comments, and an approval decision in one link.
No test manager? It can also replace the Word document, screenshot folder, and message thread used for a manual handoff.
No. Open /app and start authoring; no account is needed and manual authoring works offline once loaded. Sign in only when you choose a server feature such as sharing or sync. Browser capture uses the TestDossier Capture extension; manual authoring and screenshot paste work directly in the app.
Free authoring stays in your browser until you use a server feature. Pro syncs cases, history, and screenshots to your account. You can leave any time: export a single case, a whole project, your entire workspace with evidence, or your complete account as machine-readable files, all re-importable. Cases also export in TestRail, Xray, Zephyr, Azure DevOps, Qase, TestLink, PractiTest, and qTest shapes, plus generic CSV and Excel, for teams moving to another tool.
No. Every step and report is built deterministically from what the capture extension records, what you type, or what your CI run already describes. Your screenshots and test data are not sent to an AI, and we don't train on them.
Point your pipeline at one endpoint with a project token. Playwright, Cypress, Cucumber, JUnit, and generic JSON are read as they are; steps and failure screenshots land on the matching case, and anything unmatched is held for review rather than invented.
Free local authoring is available today. Pro is in private early access; join the early-access list and we will email you when paid access opens.
Create a report locally for free, or open the sample to see exactly what a reviewer receives.