Tom Tang
Building the evidence layer for AI-agent work.
I ship tools that make agent work inspectable, replayable, and shippable: live session dashboards, transcript archives, product-signal mining, and tracked installers.
Projects
Mission Control for AI coding agents. Observe live sessions, costs, search, sub-agents, and tool activity from one dashboard.
View project →Turn ChatGPT shared links into agent-readable archives for handoff, analysis, and future context.
Node · Codex · ClaudeMine transcripts and tickets into ranked product pain points, PRD updates, and test gaps.
Agent skill · Product opsPublic distribution layer for reusable agent tools: GitHub installs, tracked curl installers, and launch notes.
GitHub · CloudflareChat with your documents locally. No cloud, no API keys, no cost. RAG powered by Ollama.
Python · Ollama · LanceDBData-driven product comparisons. One recommendation, backed by real benchmarks.
Astro · TypeScript · PreactFive-in-a-row with an unbeatable Rust+WASM minimax AI. Runs in your browser.
Rust · WASM · React