Administration
Welcome to OctoFeedback
Sign in with your admin token. You stay signed in on this device for 30 days. Sign out on shared devices. Put Cloudflare Access in front of this console before handling sensitive feedback.
Sign in with your admin token. You stay signed in on this device for 30 days. Sign out on shared devices. Put Cloudflare Access in front of this console before handling sensitive feedback.
Want to report problems with OctoFeedback itself? Add the feedback button to this admin.
Capture reports from every project in one place.
Every report captured from your projects, newest first. Click a row to inspect it.
| Date | Project / Instance | Type | Message | Screenshot | Delivery |
|---|
Separate applications and staging/production deployments.
A project is one application whose feedback you want to collect.
An instance is one website or deployment that shows the feedback button.
Rename a project without changing its slug or registered instances. Each project groups the sites that send feedback.
When the Cloudflare API token and account id are configured in your server settings, spam-protection widgets are provisioned and synced automatically; otherwise paste the site key and secret and press Apply.
| Slug | Name | Instances | Screenshot notice | Spam protection (Turnstile) | Actions |
|---|
Origins must match the website exactly. The source security setting controls public or signed ingestion.
| Project | Instance | Allowed origin | Accepting feedback | Source security | Trash |
|---|
Choose how the feedback button looks and reads on each registered site.
Register the bundled demo project and instance for this console's address; then test annotations, uploads and routing.
Capture first; forward second. Delivery failures don't discard feedback.
A destination is another service that should receive your reports as they arrive.
Assignments override default routes for those projects. Projects without assignments use every enabled default destination, then the legacy webhook if no default exists. Disabling an assigned/default receiver pauses its deliveries and does not send them elsewhere.
Every receiver reports are delivered to, with its health at a glance.
Recent attempts and their outcome; a failed delivery can be retried on its own.
Up to 100 recent events and deliveries. Failed deliveries can be retried individually. Receivers must deduplicate X-Octo-Delivery-ID. Reports with no route remain pending until a receiver is configured.
Embed the widget and provision sources automatically.
Drop one script tag into a website, customize colors and languages, then capture screenshots.
Open integration guide ↗Manage registered sources over Ansible/REST, and use authenticated MCP tools for agent workflows.
Explore MCP and Ansible ↗Register the octofeedback-admin/production source for this exact origin and mount the widget inside the console, so you can report problems with OctoFeedback itself. Turnstile stays disabled here because the admin CSP never loads third-party scripts; sign out to remove the widget.
Revocable, optionally project-scoped credentials for Ansible and CI. Secrets appear once and must never be embedded in browser JavaScript.
| Name | Project scope | Last used | Status | Action |
|---|
OAuth 2.1 clients an administrator approved for the MCP endpoint. Revoking a grant immediately rejects its access and refresh tokens; the app must be re-authorised.
| App | Scopes | Connected | Last used | Status | Action |
|---|
Download everything — reports, screenshots, projects and settings — as one file you can keep or restore later.
Creates one file with every report, screenshot, project, instance, destination and setting.
Keep backups private. They may contain contact email, screenshots and source-token hashes. Store them encrypted. You can keep capturing feedback while the download runs.
For repeated backups from n8n, Activepieces or a scheduled task, use the authenticated paginated API or the scripts/backup-octofeedback.mjs command-line script. The browser download stops at 200 MiB; use the command-line script above that size.
Compare first, then merge (adds missing items) or replace (makes this workspace an exact copy).
Choose a backup file to compare or restore.Use the command-line script to restore files over 200 MiB.
Restoring never happens until you choose an action. Merge keeps every existing row and adds only missing ones; Replace deletes the current workspace first, so download the current data and type the confirmation phrase before it runs. A failed restore leaves maintenance mode ON until you switch it off.
Restore removed items or delete them permanently.
Items in the trash are deleted permanently after this many days. 0 keeps them until you delete them.
Removed reports you can restore or delete permanently.
Removed sites that no longer accept feedback.
Removed projects and their sites.
Removed receivers that no longer get deliveries.