Prilog bug fixing: AI bug fixing for production errors

Prilog bug fixing is AI bug fixing grounded in your production logs, traces, GitHub issues, and connected codebase. Prilog investigates the cause, prepares a patch, tests the change, and opens a pull request for human review.

From a signal to a pull request

1. Signal: connect an observability source, send OpenTelemetry logs and traces, upload a selected log file, or import a supported public GitHub issue.
2. Root cause: investigate the failure against the connected repository and the available runtime evidence.
3. Patch: prepare a code change for the affected behavior.
4. Tests: validate the proposed change with repository checks. Inspect the checks actually run and any gaps; passing checks alone do not prove a bug is fixed.
5. Pull request: publish the proposed change under your project settings so a person can review the diff, CI results, and remaining risks.

You control review and merging

Prilog never merges on its own. A human reviews every PR. Project settings control analysis, review before publication, or automatic PR creation after a policy pass. Automatic PR creation does not mean automatic merging. Read the automation controls.

Pricing and getting started

Pay $8 per publishable fix: a generated code change prepared for review and publication as a pull request. Your first 3 fixes are free, with no credit card. Codebase analysis and OpenTelemetry monitoring are free. See current pricing for the live terms.

Start with the Quickstart, explore the developer docs, and inspect public results, including upstream PRs and the separately disclosed benchmark. Public contributions do not establish elapsed fix time or imply customer endorsements.

Evaluate the workflow

Compare Sentry Seer, Datadog Bits Code, Resolve AI, Superlog, or manual coding-agent workflows. Use equivalent evidence and record tests, reviewer effort, extra edits, and unsupported cases.