On this page
AI for legal teams
Legal teams work with confidential and sensitive documents, often under professional privilege. HostYourAI runs open models on GPUs in European datacenters via vLLM, so those documents never leave the EU and processing falls under European law. You stay in control of where your data sits and who can reach it.
Why legal teams host their AI in the EU
Confidentiality and data inside Europe
All processing happens on European GPUs, with no US cloud provider in the chain. That fits the confidentiality that client files and professional privilege demand, and keeps you out of foreign access legislation such as the US CLOUD Act.
Answers with source citations
Link a knowledge base and the Router injects context from your own documents. Non-streaming requests return a source citation, so you can trace every answer back to the underlying passage instead of relying on general model knowledge.
A Data Processing Agreement and no training on your data
A Data Processing Agreement (DPA) is available and the subprocessor list is public. Your documents are not used to train models; open weights do not learn from your traffic.
What you use AI for in legal practice
Think of summarising case files, supporting contract review, legal research over your own documents and drafting correspondence. The model delivers a draft that a lawyer reviews. It is not legal advice and does not replace professional judgement.
Legal requirements and how HostYourAI meets them
| Requirement in the legal sector | How HostYourAI meets it |
|---|---|
| Confidentiality and professional privilege | Processing in the EU, isolation via dedicated instances |
| No data export outside the EU | European GPUs, no US cloud in the chain |
| Traceability of answers | Source citations on non-streaming requests via RAG |
| No reuse of client data | No training on your input or output |
| Accountability and transparency | Open models, public subprocessor list, DPA |
| Control over cost | Pay-as-you-go per token or per hour, visible per request |
How to start
Link a knowledge base with your own case files and model contracts; the Router automatically injects the relevant context, so answers are grounded in your sources. Start on the shared Router and scale to a dedicated instance for more sensitive work.
from openai import OpenAI
client = OpenAI(
base_url="https://hostyourai.com/api/v1",
api_key="hyai-...",
)
resp = client.chat.completions.create(
model="llama-3.3-70b",
messages=[{"role": "user", "content": "Summarise this case file."}],
extra_body={"knowledge_base_id": 42},
)
print(resp.choices[0].message.content)Want to know more about linking knowledge? Read the guide Build RAG on EU GPUs.
Questions about AI in the legal sector
Do confidential documents stay inside the EU?
Yes. Models run on GPUs in European datacenters and your prompts and outputs never leave the EU. There is no US cloud provider in between.
Can I trace answers back to their source?
Yes. On non-streaming requests the Router returns a source citation with document name and passage, so you can check every answer.
Are our documents used to train models?
No. There is no training on your input or output, and open weights do not learn from your traffic.
Does this replace a lawyer?
No. AI supports the groundwork and delivers a draft. A lawyer reviews and decides; it is not legal advice.
Do I get a Data Processing Agreement?
Yes, a DPA is available and the subprocessor list is public.
Can I isolate sensitive matters?
Yes. Alongside the shared Router you can run a dedicated GPU instance for isolation and predictable performance.
