Product
OrgBrain
A private LLM that answers from your organisation's own records, on hardware you control.
Design-partner programme
The problem
What's broken today
The assistants that are actually useful need your documents in someone else's cloud. For a bank, a hospital, or a public body, that ends the conversation before it starts, so years of internal policy and past decisions stay locked in folders nobody can search.
What it does
OrgBrain, in practice
OrgBrain runs an open-weight language model inside your environment and grounds it in your own documents, policies, and past decisions. Each question is resolved against the asker's existing permissions, each answer names the internal sources it came from, and each exchange is written to an audit record a compliance function can inspect.
How it works
The pipeline behind OrgBrain
Indexes your own material
OrgBrain builds its memory from the documents, policies, and records it is connected to, and that index is built and stored inside your environment.
Serves the model locally
The language model runs on GPU hardware you control. No prompt, document, or answer is sent to a third-party model provider.
Answers under your permissions
Each question is resolved against the asker's existing access rights, so OrgBrain can only draw on material that person is already cleared to read.
Writes an audit record
The question, the sources retrieved, and the answer given are logged together, so a compliance function can inspect what the system did rather than take it on trust.
Features
What OrgBrain does
Private model serving
The model weights run inside your infrastructure, so using the assistant is not a data transfer to anyone.
Organisational memory
Grounds answers in your own documents, policies, and past decisions instead of general knowledge from public training data.
Permission-aware retrieval
Honours the access rules your organisation already enforces, rather than creating a second, looser way to reach the same documents.
Source citation
Names the internal documents an answer was drawn from, so the answer can be checked against the record it came from.
Full query audit trail
Keeps an inspectable log of every question asked and every source used, which is the record a regulated review asks for first.
Open-weight model choice
Runs open-weight models selected per deployment, so the underlying model can be changed or upgraded without rebuilding the system around it.
System integrations
Connects to the document stores and internal systems the organisation already runs, so the memory stays current without a parallel upload process.
Deployment options
Where OrgBrain runs
Every option below is the same software on a different surface. The deployment model changes where it runs, not what it does.
On-premise
Model serving, index, and logs all run on hardware inside your data centre, for organisations whose rules put the boundary at the building.
Private cloud tenant
Deploys into your own cloud subscription, Azure-first, where the requirement is that data stays in an account you own rather than on hardware you own.
Air-gapped
Runs with locally held model weights and no outbound network dependency, for environments where outbound connectivity is not permitted at all.
Where it stands
Where OrgBrain stands today
Design-partner programme
OrgBrain is in its design-partner programme: the product is built and deploying with early customers while its production track record is established. It runs on the same platform as the other three products - on the customer's infrastructure, under their access controls, with nothing sent to a third-party model provider.
FAQ
Questions we get about OrgBrain
What is a private LLM?
A private LLM is a language model that runs inside an organisation's own infrastructure or cloud tenant, rather than sending prompts and data to a third-party API - built for teams where the data itself can't leave the building.
What does a self-hosted enterprise LLM mean in practice?
It means the model, the data it's trained or grounded on, and the logs of every query stay inside infrastructure your organisation controls, with your own access permissions and audit trail, not a vendor's.
What is an LLM with organisational memory?
It's a model connected to your organisation's own knowledge - documents, policies, past decisions - so it can answer questions about your business specifically, instead of only general knowledge from public training data.
How does OrgBrain support data sovereignty requirements?
OrgBrain is deployed on-premise or inside your own cloud tenant, so no prompt, document, or output has to leave your environment - the starting requirement for teams working under data sovereignty rules, not an add-on.
Who is OrgBrain built for right now?
OrgBrain is built for regulated organisations that need a private, auditable assistant and are willing to work closely with us as an early design partner while the product matures.
Ready to see OrgBrain for yourself?
A private LLM that answers from your organisation's own records, on hardware you control.