LangWatch adds session replay, tool-call tracking, and cost visibility for coding agents
_rchaves_ · x · 2026-07-29
The post introduces npx langwatch claude, a free-for-individuals observability tool for AI coding-agent sessions.
It tracks the actual cost of each session, cache reads and writes, every tool call, token usage, full terminal replay, and the exact moment the agent started going off the rails. The author says it also works with OpenAI Codex, Gemini CLI, and OpenCode, and argues that watching agent behavior should become standard before the tokens eat the budget.
Related event: LangWatch Introduces Usage Tracking for Claude Code(2 posts)→
More from coding & agent
- Running Claude Code on Mobile Ends Doomscrolling; PAL Tool Syncs Memory Across Devices — rchardkovacs · 2026-07-30
- Dev Kills Complex AI Agent Frameworks with 150 Lines of Code and a Cron Job — Warm-Reaction-456 · 2026-07-30
- AI Security Benchmarks: First Task the Agent to Break Out of the Sandbox — j_foerst · 2026-07-30
- Ontologies Are Back: UC Berkeley Professor Explains How They Guardrail AI Agents — Latent Space · 2026-07-30
- GitHub Copilot CLI Adds Multi-Model Cross-Review, Closing 74.7% Performance Gap — 0xkarasy · 2026-07-30
- AI Agents Without Stop Policies Become Infinite Loops: Three Gates for Production — ArtComprehensive7403 · 2026-07-30