Add Cloudglue Session Recap template with browse-and-recap action#142
Closed
kdr wants to merge 2 commits intokernel:mainfrom
Closed
Add Cloudglue Session Recap template with browse-and-recap action#142kdr wants to merge 2 commits intokernel:mainfrom
kdr wants to merge 2 commits intokernel:mainfrom
Conversation
Add a new TypeScript template that analyzes video recordings using Cloudglue's describe and segment-level extract APIs. Takes any video URL (Kernel session recording, screen capture, or video file) and produces a structured scene-by-scene recap with thumbnails, timestamps, user actions, screen descriptions, and a complete markdown document. - Uses adaptive shot detection with configurable max_seconds (2-60, default 8) - Runs describe and extract in parallel for faster results - Generates thumbnail preview grid + detailed scene breakdown - Deduplicates scenes with matching timestamps from describe/extract misalignment - Registered in Go template system with deploy config Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add `browse-and-recap` action: opens a Kernel browser, records the session, then analyzes the replay with Cloudglue. Includes placeholder for custom browser automation logic. - Add retry with 20s backoff on describe/extract API calls (2 attempts each, independent retries) for cases where video isn't ready yet - Remove url_at_step from extract schema (was hallucinating URLs) - Improve markdown output: bullet-pointed metadata, thumbnail preview rename, generated title as subheader in summary section - Deduplicate scenes with matching timestamps from describe/extract segment count misalignment - Extract `analyzeRecording` as shared function for both actions - Add playwright-core dependency for browse-and-recap Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
cloudglue-session-recapTypeScript template that analyzes video recordings using Cloudglue APIssession-recap— takes any video URL and produces a structured scene-by-scene recapbrowse-and-recap— opens a Kernel browser, records the session, then analyzes the replay with Cloudglue (demo action with placeholder for custom browser automation)Example breakdown for a scrolling through of hacker news in this video https://media.cloudglue.dev/test-data/browse-hackernews.mp4
Details
max_seconds(2-60, default 8)pkg/create/templates.go)Test plan
make buildpassesmake testpasses (all Go tests green)make lint— all 68 issues are pre-existing (none in our changes, which are TypeScript template files + Go template registry)kernel create --template cloudglue-session-recap --language typescriptscaffolds correctlynpx tsc --noEmit)_gitignorecorrectly renamed to.gitignoreduring scaffolding.env.examplecopied correctlypnpm installsucceeds in scaffolded directorykernel deploy index.ts --env-file .envdeploys successfullysession-recaptested with multiple video URLs (screen recordings, food videos)browse-and-recaptested with live Kernel browser sessions🤖 Generated with Claude Code