In-App Feedback
Capture bugs and feature requests in the moment — while a user is actually using your app — with screen recording, voice narration, and annotated screenshots, then triage them on a board in your dashboard.
How it works
- Capture. A drop-in widget lets the user record their screen while narrating with their voice, snap and draw on a screenshot, choose bug / feature / question, and submit. See the JavaScript SDK guide for installation.
- Upload. Attachments upload directly to private storage via signed URLs — nothing large passes through your servers.
- AI enrichment. Gemini transcribes the voice narration and drafts a title, type, priority and summary from what the user described. Any field the user filled in themselves is kept; the AI only fills the blanks. This is credit-metered.
- Triage. Items appear on the Feedback board in your dashboard.
- Build. A Codex agent using
@simplr-ai/dev-mcpcan list, inspect and claim a ticket, then attach its pull request and progress comments. Identified submitters receive a clear email as the ticket moves through each stage. - Release. After deployment, the agent marks the ticket released. Simplr closes it and sends the final release email, including the public release URL when one is provided.
The triage board
Open Dashboard → Feedback. Each card shows the type, priority, a screenshot thumbnail and a recording indicator. Open one to play the recording, read the transcript and AI summary, view annotated screenshots, see the page URL and RUM session it came from, and leave comments.
Items move through an active pipeline:
New → In Progress → Testable → Done
Testable is an explicit column so QA knows what is ready to verify. You can also park items that aren't going into the pipeline:
- Wishlist — nice-to-have, revisit later
- Backlog — not prioritized right now
- Won't Do — declined
- Duplicate — already tracked elsewhere
Every status change, assignment, priority change and comment is recorded on the item's history.
Connect Codex to the feedback queue
Configure @simplr-ai/dev-mcp with the Simplr API URL and a secret organization key:
claude mcp add simplr-dev \
-e SIMPLR_API_URL=https://api.simplr-ai.com \
-e SIMPLR_API_KEY=sk_live_xxx \
-- npx -y @simplr-ai/dev-mcp
The coding workflow uses these tools:
list_feedback → get_feedback → claim_feedback → code and test → update_feedback / comment_feedback → deploy → release_feedback
Identified submitters receive deduplicated lifecycle emails when feedback is received, moves into progress, becomes testable, or reaches any terminal/parked outcome. release_feedback sends the final released email and should be called only after a real deployment, with an optional public release-note URL.
What gets captured
Along with the recording, audio and screenshots, each item stores the page URL (query string stripped by default), device context (browser, OS, viewport, language, timezone), the reporter's application user id/email/name if provided, and the linked RUM session id when RUM is running — so you can jump straight to what the user was doing.