TraceX reads a customer's email, Slack message or Jira ticket, investigates your database, code and logs read-only, and returns the root cause with evidence for every claim, a fix as a pull request and a drafted reply. Free in private beta.
Every engineering team knows this message: "My refund never arrived." Someone stops what they're doing, opens the database, greps the code, reads last week's deploys, checks Sentry, and an afternoon later finds that a migration renamed a state the refund worker still filters on.
We built TraceX to do that investigation for you.
How it works: an issue comes in by email, Slack, Jira or a question in the dashboard. TraceX plans its checks and works through your database, code and logs, the way your best engineer would. You watch every step as it happens. It comes back with:
• the root cause, with its confidence and the evidence for every claim, and the hypotheses it ruled out • the fix as exact code edits, opened as a pull request in one click • a reply to the customer, drafted with the facts
What we cared about most:
• Read-only. Database connections use read-only credentials, write queries are refused before they run, and code changes only ever arrive as pull requests you review. • Nothing kept. TraceX reads your systems live and keeps a note of what it looked up, not your rows. It never trains on your data. • Answers you can check. A second model checks every answer against the evidence before you see it.
It also works as a chat for your whole team ("how many refunds are stuck, and since when?"), reviews pull requests with your codebase as context, and has a CLI that answers in your terminal and works as a coding agent that asks before every change.
The evidence-for-every-claim part is what would sell this to me. A root-cause answer I can't check is just another confident guess, and the ruled-out hypotheses are the bit I'd actually want in front of a customer — it shows the work instead of just the verdict. Curious how it behaves when the evidence points two ways, or when the fix it proposes touches a migration — does it still draft the PR, or hold back and ask?
read-only creds is good but on prod it's not only about writes. we had an incident where our pool was saturated, the last thing I want then is an agent sending more queries on the primary. can I point it to a replica or a Neon branch?
About TraceX on Product Hunt
“The AI engineer for your production issues”
TraceX launched on Product Hunt on November 6th, 2026 and earned 65 upvotes and 6 comments, placing #11 on the daily leaderboard. TraceX reads a customer's email, Slack message or Jira ticket, investigates your database, code and logs read-only, and returns the root cause with evidence for every claim, a fix as a pull request and a drafted reply. Free in private beta.
TraceX was featured in SaaS (45.2k followers), Developer Tools (522.6k followers) and Artificial Intelligence (483k followers) on Product Hunt. Together, these topics include over 288.9k products, making this a competitive space to launch in.
Who hunted TraceX?
TraceX was hunted by Shashank Saraswat. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how TraceX stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Shashank Saraswat
Hey Product Hunt 👋
Every engineering team knows this message: "My refund never arrived." Someone stops what they're doing, opens the database, greps the code, reads last week's deploys, checks Sentry, and an afternoon later finds that a migration renamed a state the refund worker still filters on.
We built TraceX to do that investigation for you.
How it works: an issue comes in by email, Slack, Jira or a question in the dashboard. TraceX plans its checks and works through your database, code and logs, the way your best engineer would. You watch every step as it happens. It comes back with:
• the root cause, with its confidence and the evidence for every claim, and the hypotheses it ruled out • the fix as exact code edits, opened as a pull request in one click • a reply to the customer, drafted with the facts
What we cared about most:
• Read-only. Database connections use read-only credentials, write queries are refused before they run, and code changes only ever arrive as pull requests you review. • Nothing kept. TraceX reads your systems live and keeps a note of what it looked up, not your rows. It never trains on your data. • Answers you can check. A second model checks every answer against the evidence before you see it.
It also works as a chat for your whole team ("how many refunds are stuck, and since when?"), reviews pull requests with your codebase as context, and has a CLI that answers in your terminal and works as a coding agent that asks before every change.
It's free while we're in private beta. You can try the sandbox without signing up: https://app.tracexpro.com/sandbox
I'd love your honest feedback, especially on where TraceX gets it wrong. I'll be here all day answering questions.