I build practical tools where AI, marketing operations, and automation meet.

I care about software that earns its place in a workflow: small enough to understand, clear enough to audit, and useful before a team adopts a large platform. Most of the work here is local-first by default, with explicit rules and evidence rather than hidden automation.

Evidence before automation. A useful system should retain the prompt, source data, or rule behind a decision.
Local first. The examples run without handing customer data to a third party.
Small, composable tools. I would rather build a sharp utility that fits into an existing workflow than an unfinished all-in-one platform.
Honest boundaries. Reports describe what the supplied data says; they do not make guarantees about rank, conversion, or revenue.

Every Friday at 18:30 Asia/Tehran, this portfolio is reviewed for one meaningful, testable improvement. A new repository is only released when it solves a clear problem and includes runnable code, human documentation, a license, safe example data, and a quality check. The goal is a useful public record of work—not artificial activity.

Looking for a starting point? Try the AnswerTrace live demo, then use its repository for the local development path.

The changing shape of AI search, practical analytics hygiene, and systems that give small marketing teams a transparent way to review their data and decisions.

Found a confusing edge case, an unclear README, or a more useful workflow? Open an issue with a small reproducible example. I value practical feedback more than vague feature lists.

_{Build the signal. Keep the trail.}