Comparison

Where Lowkey has the edge.

For brainstorming or summarising public text, a general AI chatbot is strong, and a broad legal copilot is valuable as a research and writing assistant. Lowkey is made for a different moment: the form work of a Dutch tax procedure, with client data, finished into a form with a cover letter. That is where our edge sits. Here is what it looks like.

  General AI chatbot Broad legal copilot Lowkey
Made for Broad text and questions Research and drafts across the board One Dutch tax procedure, finished
End product Text and answers A memo or draft Filled form plus cover letter
Output Free text Free text Deterministic: same input, same result
Client data Varies by provider Varies by provider EU only, processed in memory and discarded
Pricing model Per-user subscription Contract and fixed seats EUR 95 per finished document, no seats

One procedure, fully finished

Lowkey does one thing: the form work of a Dutch tax procedure, from the data to a filled form in the official tax format with a cover letter on your own letterhead. All you do is check it, sign it and file it.

An output you can build on

The math and form-filling are deterministic code, so the same input gives the same result every time. A model at most reads a messy upload so you do not retype anything. It never decides the numbers.

EU only and nothing stored

Every layer that touches your client data sits in the EU, and uploads are processed in memory and then discarded. Your firm stays the controller, Lowkey is your processor. That fits your professional confidentiality and the GDPR. Read how we handle data.

Pay per finished document

No seats and no subscription: you pay EUR 95 per finished document. For an SME firm that handles a 30% ruling or a fiscal unity a few times a month, that matches how often you actually use it.

We name no brands here. This is about the categories: general AI assistants and broad legal copilots. All useful in their own area. This page shows where Lowkey is focused.

Create account EU & data