Question 1

How do I connect my codebase, database and logs to an AI agent?

Accepted Answer

TraceX connects to each source through a read-only integration you approve per system. Codebases connect through GitHub, GitLab, Bitbucket or Azure DevOps; databases through PostgreSQL, MySQL, MongoDB or Snowflake; logs and monitoring through Datadog, Grafana, Splunk or CloudWatch; mail through Gmail or Outlook; tickets through Jira, Linear, ServiceNow or Zendesk. Once connected, the agent reads across all of them in a single investigation instead of one tool at a time. Access is read-only by default and can be revoked at any time.

Question 2

How can AI debug a production issue automatically?

Accepted Answer

TraceX opens an investigation from the incoming signal, such as a customer email or a ticket, and then reads what a senior engineer would read: recent code changes, the affected database records, service logs around the failure window, and prior investigations of the same area. It correlates them into a root cause, cites the exact evidence it used, and drafts both the code fix and the reply to the customer. A worked example: 1,204 payments failed after a release turned off automatic retries, so any charge the bank took longer than three seconds to confirm failed instead of retrying.

Question 3

Can non-engineers ask questions about company systems in plain English?

Accepted Answer

Yes. The TraceX Console lets support, operations, finance and product ask across code, databases, logs and mail in plain English, with no SQL, no dashboards and no access requests. Every answer links to the sources it read, so it can be checked rather than trusted.

Question 4

Is it safe to connect production systems to an AI investigator?

Accepted Answer

TraceX is read-only by default and never writes to a connected system unless that is explicitly granted. Customer code, records and logs are never used to train foundation models. It supports SSO, role-scoped access, full audit logs and end-to-end encryption, and access to any source can be revoked at any time.

Question 5

How is TraceX different from an AI coding assistant or a chatbot?

Accepted Answer

A coding assistant works inside one repository and helps write code. TraceX investigates a problem across an entire ecosystem — mail, tickets, databases, logs, monitoring, infrastructure, documentation and source — the way a second- or third-line engineer does, and every completed investigation becomes organizational memory the next one can draw on.

Question 6

How much does TraceX cost?

Accepted Answer

The Developer plan is 39 US dollars per developer per month and includes 3 connected sources and 100 investigations a month. The Project plan is 390 US dollars per project per month and includes up to 10 seats, unlimited connected sources, 2,000 investigations a month, SSO and full audit logs. Enterprise pricing is scoped per organization and covers private or self-hosted tenancy, data residency controls, SCIM and SLAs.

TraceX — the AI investigation platform for your whole stack

Connect your codebase, databases, logs and tickets to AI

How to debug a production issue with AI

Safe to point at production

Pricing