Kolo

Kolo's core idea is simple and it is the right one: capture everything as text while the code runs, then let developers query it afterward. The always-on tracer records every function call, return value, local variable, and Django HTTP request, and the web UI turns that into an execution graph you can navigate. The greppable text trace format is the underrated part — it means Kolo output behaves like a log file you can grep, diff, and export, rather than an opaque binary capture. The standout capability remains test generation: Kolo takes a live runtime trace and produces working Django integration tests, which removes a real chunk of manual boilerplate for views and models. The playground lowers the trial cost to zero — you can see a trace without installing the package. Where Kolo is honest about its edges: it is Django-first. The vendor guidance points at Django request tracing and Django integration tests, while broader Python support is framed as in development. Non-Django frameworks such as Flask and FastAPI are not where this tool is aimed today. It is also not an APM or error-tracking replacement — there is nothing here about production alerting, uptime, or error aggregation, so teams looking for Sentry- or Datadog-style monitoring should keep those tools and use Kolo during development. Collaboration features are listed on the vendor site but the workflow is fundamentally a local developer tool. If you fit the profile — Python, Django, debugging complex request flows, or a test backlog you have been avoiding — Kolo is one of the more distinct free tools in this space.