Glassbrain

Glassbrain is a visual debugging engine for AI-powered applications that captures every LLM call, tool invocation, and chain step in your app. It displays your AI's entire reasoning chain as an interactive, clickable trace tree so you can see exactly where and why something went wrong. When a bug is found, you can click any node, swap inputs, and replay from that point without redeploying. Setup takes under five minutes with a single npm or pip install.

• Visual Trace Tree — See every LLM call, tool use, and chain step in a collapsible, interactive tree view for instant situational awareness.
• Time-Travel Replay — Click any node in the trace, change the input, and replay execution from that point forward without redeploying your application.
• AI Fix Suggestions — Get concrete, actionable fix proposals powered by Claude, including prompt changes, temperature adjustments, and configuration tweaks.
• Diff View — Compare two traces side by side to see exactly what changed between a broken and a fixed execution.
• Shareable Links — Generate a URL for any debugging session and share it with teammates so they see the full trace context.
• One-Line Setup — Install the SDK with npm install @glassbrain/js or pip install glassbrain, call init() once, and start capturing traces automatically.
• Broad Framework Support — Works with OpenAI, Anthropic, LangChain, LlamaIndex, and any custom stack via an OpenTelemetry-compatible endpoint.
• Team Collaboration — Higher-tier plans include team members, shareable replay links, Slack/Discord alerts, and CI/CD integration.