Query session records and build talk slides with Claude Code
A CS instructor used Claude Code to query his recorded session database for commit stats and timelines and to build slides for a faculty talk.
Done withClaude Code
- What they did
- Geoffrey Challen had built Claude Recorder, a tool that logs his Claude Code sessions to a local SQLite database with timestamps, tool use and token counts. For his faculty retreat talk, he had Claude Code query that database for stats and commit timelines. It then designed the reveal.js slides, including an interactive timeline, as one standalone HTML file. It also transcribed his practice recordings with WhisperX and drafted the written summary.
- How it went
- The stats covered one 17-hour day: 903 prompts, 63 commits, and about 16,000 lines added. Claude produced the slides and the write-up. The source doesn't describe any failures in the slide-building itself.
- Worth knowing
- The session data comes from a custom recorder he built beforehand. The day's API usage would have cost about $1,000, but he's on the $200 per month Max plan.
Try it yourself with Claude Code
I have a database of recorded work sessions at [path to database]. Explore its structure, then answer these questions: [e.g. commit counts per week, a timeline of the project, the biggest changes]. Then build a slide deck for a [length]-minute talk to [audience] on [topic] using those findings. You're done when I have the numbers with a note on how you computed them, plus a slide file I can open. Don't modify the database.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: geoffreychallen.com · Jan 15, 2026
Related use cases
Turn tickets into merged pull requests
Pulls tasks from GitHub Issues or Linear, then runs Claude Code or Codex in isolated Kubernetes pods to carry each one through to a PR.
Audit subscriptions and email companies to cancel them
Claude Code pulled transactions across credit cards, flagged recurring charges, and drafted cancellation emails, recovering $700.
Audit ad accounts across six platforms
A Claude Code skill runs parallel subagents performing 190 audit checks across ad platforms and produces a prioritized action plan.
Run persistent financial research with an agent
LangAlpha is a Claude Code-style agent for investing research that keeps persistent workspaces and memory so analysis such as models and comps can be updated across sessions.
Custom agent
Try Devin on research tasks (mostly failed)
Answer.AI had Devin fetch an author's recent papers and research transcript summarization; it got stuck parsing HTML and returned generic results, rated a failure.
Run academic literature reviews from seed papers
Poster built an agent skill that starts from seed papers, searches backward and forward, screens candidates, and iteratively refines a literature review.
Custom agent
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.