Vahtian

I build the tools my team and I need

I’m Heidi Andersén, MD, PhD, a respiratory physician and researcher in Finland. My work spans respiratory medicine, thoracic oncology, and clinical research methodology. Vahtian grew from the problems my team and I kept solving over and over again.

The guides in Learn are written for master’s and doctoral students in medicine, and for the supervisors who read their drafts. Little of it is medicine-specific.

The work

Most of my days are spent designing studies, analysing data, reviewing evidence, and writing manuscripts. The same problems kept returning: claims no longer matched their sources, important decisions were never documented, and sensitive data could not leave the clinical environment.

A research decision may need explaining months or years later. The record has to exist before the question comes.

March asks about September.

When a reviewer asks in March about a choice I made in September, I need the record, not my memory. The tools are built for that day.

Born out of frustration

Every tool was built because my team or I needed it first.

  • A REDCap data dictionary that refused to validate became DictVahti.
  • Checking a reference list by hand became CiteVahti.

Everything I build starts inside active research before it becomes software, and it is tested on my own manuscripts long before anyone else downloads it.

Built for my team, shared with yours XOXO

The checking tools are free and run in your browser or on your machine. The core methods are open source. The code is there to inspect, test and reuse. The paid kits package the workflows my team uses, with the templates and records included.

AI should help researchers.

It can search, compare, summarise and suggest. It should never quietly become the scientific judgement. Every important decision should still make sense months or years later.

How the software is designed

Technology changes quickly. Good research principles do not: evidence stays connected to every important claim, and scientific work gets stress-tested before it is trusted. So the rules are few.

  1. You decide

    AI assists; it does not decide. Researchers make the final call, every time.

  2. The record stays

    Every tool leaves a trail you can inspect, share, and re-run.

  3. Data stays under your control

    The software runs where your data already lives whenever the workflow allows. Your organisation keeps control of its data.

  4. Checks stay transparent

    Every assessment explains what was checked and why it was flagged.

Features change. These principles do not.

vahti, Finnish for sentinel / guard.

Find the work

The tools grow from feedback. If something helps, or something fights you, say so.