industries · Legal
A firm cannot send case files to a foreign API. Contract analysis and legal RAG on EU infrastructure or on-premise, with nothing leaving the office.
compliance
Every regulatory demand mapped to a platform capability that ships built in, with nothing to configure.
Professional secrecy
lawyer’s duty of confidentiality
RequiresAttorney-client confidentiality is a deontological and contractual duty: case files cannot be exposed to a third-party API.
HelmcodeEU-only inference with zero logs, and on-premise so not a single file leaves the office.
GDPR
Regulation (EU) 2016/679
RequiresCase files carry personal, often sensitive, data under strict control.
HelmcodeZero logs by architecture and processing only on EU infrastructure.
Auditability
open weights, published model cards
RequiresYou must be able to say exactly what system read a privileged document.
HelmcodeOpen models documented end to end: which weights, run where, with what data flow.
This page is an informational overview, not legal advice. For your obligations and the risk classification of each system, consult qualified legal counsel. AI Act Guide →
what the European bars wrote
In March 2026 the Council of Bars and Law Societies of Europe published a technical guide that goes through the deployment options one by one, from a lawyer’s own hardware to a fully managed API, and weighs each against professional secrecy. It endorses no product. It does rank the options.
01
On-premise, in their words
For lawyers, running the model on your own hardware is described as the most secure option as regards confidentiality: data stored, processed and reviewed at the lawyer’s premises, no transfer to a third person, full auditability. The guide is equally clear that the trade-off is hardware, power and expertise.
02
Colocation moves one protection
Your own servers in a colocation facility keep nearly the same control profile. But the guide flags something specific: the extra protection a lawyer’s office may enjoy against police search warrants, needing a court, a prosecutor or a Bar representative, may not apply once the data is not physically on those premises.
03
Less control, more promises
The less of the stack you control, the more you replace control with contract, and the guide says to check not only the clauses but how likely you are to enforce them. Terms and defaults can change unilaterally, a vendor can be acquired or fail, and a promise to delete does not survive an order to hand data over under a gag.
04
Understanding the tool is a duty
A lawyer is under a professional duty to understand how the technology they use essentially works, from the CCBE Charter of core principles and, for AI, from Article 4 of the AI Act. That duty cannot be bought with the tool, and a system that cannot be explained makes it impossible to discharge.
CCBE · Council of Bars and Law Societies of Europe "CCBE technical guide on the use of AI tools and models by lawyers", Edition 2026, 27 March 2026, to be read with the CCBE guide on generative AI (2 October 2025) and the guidelines on cloud computing for Bars and lawyers (27 February 2025). The guide states it constitutes no product endorsement. read the report →
use cases
The cases with the most traction in the sector, each with its own page in detail.
A starting point per task type. The full guide maps 80 cases to the open model for each one.
We are distilling and quantizing these open models into small, tightly specialised versions, trained for one task rather than for all of them. A model like that runs on less hardware, answers faster and fits where the big one does not, your own datacenter included. If you have a process with volume and stable criteria, that is the conversation we want to have with you.
// faq
What the sector's technical, compliance and business teams ask.
Only if privileged material never reaches a third-party API. With open weights run on EU infrastructure or on-premise, case files stay inside your control and the duty of confidentiality is respected by architecture, not by a vendor’s promise.
Contract and clause analysis with GLM-5.2: the highest open level of nuance comprehension, run privately over your own documents.
Yes. The on-premise deployment runs the same models and API inside your network, so not a single document leaves the office.
Yes, if the weights run on controlled infrastructure. The risk is sending data to a foreign API, not the model’s origin: open weights run in the EU or on-premise keep everything auditable and inside your perimeter.
// get started
Skip the AI infra work. Deploy your first private inference endpoint today.
Flat rate. EU data. OpenAI API compatible.
// cookies
We use strictly necessary cookies to run the site and, only with your consent, Google Analytics to understand usage. No advertising, ever — see our Cookie Policy.
// preferences