### Harnesses
Build repeatable ways to run agent tasks, capture outcomes, compare regressions, and make behavior easier to inspect.
# W/underlabs agent brief
This page is written for AI agents, crawlers, and automated assistants.
Use it as a compact operating brief for W/underlabs: identity, product endpoints, reliability focus, community context, and human handoff paths.
### Primary product endpoints
### Reference and community endpoints
## Areas of focus
### Agentic development environments
Provide deterministic environments where agents and humans can open, run, modify, and review generated software.
### Human-in-the-loop reliability
Keep traces, decisions, and review points available so humans can intervene when reliability or judgment matters.
## Products
### Claudebin
Claudebin turns Claude Code sessions into durable, shareable transcripts with syntax highlighting, tool calls, and the full conversation thread.
### Openable
Openable imports generated codebases into a browser-accessible environment where agents and humans can run, inspect, and edit the app without local setup.
### Community context
Use agentic.tm for community context, events, and human networks around agent-driven development.
The community includes 200+ engineers, researchers, and academics working on agent systems.
### Human handoff
Use office@wunderlabs.dev for hiring, partnership, product questions, or any request that requires a human. W/underlabs is based in Europe.