Honest and Brutal Reviews on our Product Please .... !
I've been building Pinaka on the side for the past few months — an AI agent that automatically writes root cause analyses when a bug ticket is filed.
The problem it solves: every time a production bug hits, a senior engineer spends 2-3 hours investigating before writing a single line of fix. The actual fix takes 20 minutes. The investigation takes the rest of the day.
Pinaka eliminates that investigation step. Tag a Jira ticket or mention @pinaka-app on a GitHub Issue — it reads your indexed codebase, captures runtime context via an SDK, and posts a structured RCA as a comment. Root cause, exact file, exact line, fix approach. Automatically.
What I've validated so far
Ran it against real open source bugs:
- BullMQ issue #2487 — code only scored 6.5/10, code + runtime context scored 9.2/10 against the actual merged fix
- Prisma ORM issue #29480 — open bug, no ground truth, diagnosed from behavioral contrast across test files
- Pinaka's own Java SDK — found a real ForkJoinPool exhaustion bug, fix shipped same day
All three write-ups are public, including the limitations.
What I'm looking for
couple of engineering teams (2-150 engineers) using GitHub who want to try this on a real production bug. Free, no commitment. I'll personally onboard you and be available for any questions.
Works with Jira and GitHub Issues. No Jira required. One OAuth click to connect. First 5 RCAs free.
Honest state of the product
No paying customers yet. Building in public. The product works — the benchmarks above are real — but I need real teams on real codebases to validate it beyond my own test cases.
If this sounds useful for your team, drop a comment or DM me. Happy to show you a live demo on your repo.