HomeCustomer storiesThe R&D centre of a SaaS vendor
Software services · Customer story

Debugging sessions became team assets, cutting integration time in half

The R&D centre of a SaaS vendor

Delivery outcomes
50%+
average drop in integration time
100%
of production issues reproducible
3 shells
sharing the same kernel
JSON
default output for AI consumption
Timeline3 weeks (tool rollout and process change)

What it looked like before

Front-end and back-end integration had long run on “try it again on your side”. Mobile issues were the hardest to localise: client logs were incomplete, intermittent production faults resisted reproduction, and a single API issue took two to three days of back-and-forth on average. The team also wanted to bring AI into triage but lacked structured traffic data.

What we delivered

pxc became the unified capture and debugging tool: a Go kernel with macOS / Windows shells, breakpoints that rewrite requests to test a hypothesis, and full HAR exports attached to issues. The CLI emits JSON for AI agents to consume, folding traffic analysis into automated triage.

  • Suspend at a breakpoint and rewrite the request to test a hypothesis — no interception scripts, no code changes and redeployment
  • Full HAR exports attached to issues let a colleague reproduce the problem without asking you to “try it once more”
  • Mobile capture filled in the client-side real requests that had been the hardest to obtain
  • Structured CLI output lets AI read the traffic and locate the problem directly, instead of a human describing screenshots

The most valuable thing is not the capture itself — it is that a debugging session can be shared. Problems no longer travel with the person.

—— Head of engineering effectiveness
More Stories
Get in touch

Hand the complexity of identity, agents and private domain to one governable kernel

Whether you are replacing an existing IAM, building an agent platform, or trying to make private-domain operations actually work — start with a 30-minute architecture call. We will first judge whether this is the kind of problem we are good at, and say so plainly if it is not.