Tuesday, August 11, 2026
5:00 PM – 7:30 PM CDT
AI That Works: Build an Apify Actor That Reads Your Claude Code Logs
About this event
A live-build evening in Austin: turn your own Claude Code session logs into a structured map of the tools, MCP servers and packages you actually use — and the gaps you don't yet have.
Reading your own transcripts
AI That Works is a monthly Austin meetup for people putting AI to work on something real — developers, founders, operators, go-to-market teams, and the AI-curious. Each session pairs live demos from people shipping in production with a hands-on build and a candid conversation about what is actually producing results.
This month's build starts from an observation: you probably think you know which AI tools you lean on and which you are missing, but the only honest record is the session log sitting on your own disk, and almost nobody reads it.
What gets built
The session works through a self-improving research loop in stages:
- The capture layer — an Apify Actor that parses Claude Code JSONL session files and extracts a structured usage taxonomy: tool-use events, MCP server names, imported packages, and the domains touched
- A scoring model that cross-references that taxonomy against what you already have installed and ranks the gaps
- A webhook loop that turns a finished run into fresh research the moment it exists
You leave with the working stack, not just a demo. The organizers are explicit that it does not matter whether you write code or mostly prompt.
Practical details
- When: Tuesday, August 11, 5:00–7:30 p.m.
- Where: Antler VC, Austin
- Cost: Free; RSVP through Luma
Organized by
A platform for building and running "Actors" — deployable cloud programs for web scraping, data extraction, and agent tooling. Its community team runs the monthly Austin "AI That Works" meetup, a live-build session for developers, founders, and operators shipping real AI workflows.