This is Seamlift’s own verification record. We used generated example data in a local development environment to check the claims on our video editor page. The results below distinguish exercised behavior from a review of the implementation.
These checks are not an independent review or a test of the live service. The illustrated project and comments on the landing page are examples; they are not customer testimonials or measured results.
In the local preview, a signed-out guest played a generated 35-second clip, clicked a comment at 00:18, and saw the player seek to 18 seconds. We posted a new timestamped comment, reloaded the page, and confirmed that it remained. The v2 link showed a separate, empty comment history.
The browser-posted note is the third comment; the first two were created by the API verification. This used the current review UI with local media and the local backend, without exercising uploads, transcoding, or production delivery. Read the browser observations and setup.
We exercised the running review API with a real local database. Guest requests carried no account credentials. All generated test records were removed afterward.
A new user record defaulted to Free and resolved to 10 GB, with no subscription or payment record created. This checks the account defaults, not the complete signup flow.
See plan options and the signup form. Confirm paid prices at checkout before upgrading.
The running public API returned the shared version with HTTP 200 to a request with no cookies or Authorization header. Video playback was not part of this API check.
A request with no account credentials saved a comment at 18 seconds, returned HTTP 201, and the comment list returned the same timestamp. The guest received an assigned name and avatar.
Two comments with the same visitor identifier on one version received the same assigned guest name. The request supplied no name or email. A visitor identifier is stored, so this is not a claim of untraceable anonymity.
A link created with the default setting opened successfully but rejected a guest comment with HTTP 403. It could still read the version’s existing comments; disabling guest posting does not hide feedback.
With v1 and v2 in the same project, the original link still returned v1. The v2 link returned v2 with no v1 comments. A second link to v1 could read v1’s existing comments.
After invoking the product’s revoke handler, the running public API rejected version access (410), comment reads (404), and new comments (404). Another link remained active. This does not test media URLs, already-open sessions, downloads, or the owner authorization layer.
These claims have a narrower evidence basis than a completed end-to-end product test.
The project-creation path has no plan-based project-count check. The current Free plan provides 10 GB of base storage. Unlimited projects does not mean unlimited storage, and we have not load-tested an unlimited number of projects.
The signup interface offers Google sign-in and an email form with name and password fields. It has no payment form, and the account-creation path assigns the Free plan without creating a paid subscription. Completing signup was not part of this verification run.
The project interface and server include upload, version-list, and download actions. The verification used generated version records, so it did not test file transfer, transcoding, playback compatibility, or download completion.
We have not measured time saved, faster approvals, improved conversion, or performance against Frame.io. This record also does not verify YouTube publishing, production uptime, media-access security, or compatibility across browsers. Those claims need their own tests or attributable customer evidence.