ownedai

Services

Two services are available today, both fixed-scope and delivered on your hardware. Pricing on request. What is still in development is listed under Projects.

Private LLM stack installation

A private, self-hosted language-model environment on a server you own or specify. Open-weight models, local inference, a chat interface for your team. Vector store, observability and database are added only when an application needs them. No API keys, no external calls, no data leaving the building.

Ask about an installation

Who it is for

Law firms, financial services, healthcare providers, public bodies and SMEs that handle confidential material and cannot send it to a cloud AI provider.

Typical timeline and price

One to two weeks from access to handover. Fixed scope; price on request.

What you get

  • Running stack on your hardware, or on hardware we specify for you
  • Installation record: what was installed, which versions, how it was configured
  • Hardening checklist: access, network isolation, updates, backups
  • Handover session for your IT and first users
  • Written statement of what the system can and cannot do

What you need

  • A Linux server with a suitable GPU, or a budget for one — we advise on specification
  • One IT contact for access and network questions
  • Half a day of your time for handover

What we decide together

  • Which open-weight model, and at what size, for your hardware and your documents
  • Your existing server, or a specification we write for you
  • Interface for your team, and who administers it
  • Whether the optional layer (vector store, observability) is needed at all

What it is not

Not a cloud subscription, not a managed service, not a validated GxP system. A private environment you own and run, with a written statement of what it can and cannot do.

Reference deployment

Reference deployment: every component inside the client perimeter, no outbound connections your server · your perimeter GPUlocal inference Ollamaopen-weight model Open WebUIchat for your team Qdrantvector store Langfuseobservability PostgreSQLfor Langfuse Docker Compose no outboundconnections Core — always installed Optional — added only when an application needs them

OwnedPulse runs as one more container on the same base, and is the first application that uses the optional layer.

OwnedPulse deployment

OwnedPulse is our open-source regulatory intelligence tool. It monitors regulatory sources you choose, retrieves the documents, and answers questions grounded in those documents — with the source shown for every answer.

We install and configure it on your infrastructure, tuned to the sources and topics your regulatory affairs or quality function actually follows. Screenshots, the traceability drawer and the audit export are shown on the OwnedPulse page.

Ask about OwnedPulse

Who it is for

Regulatory affairs, quality assurance and compliance teams in pharma, medical devices, CROs and CDMOs who spend hours each week tracking EMA, FDA and national guidance.

Typical timeline and price

One to two weeks. Fixed scope; price on request.

What you get

  • OwnedPulse deployed on your server
  • Source configuration for your regulators and topics
  • Scheduled updates
  • Written context-of-use statement: what the tool is for, what it is not for
  • Thirty days of support after go-live

What you need

  • A Linux server with GPU, or the same stack as the installation above
  • A list of the regulatory sources and topics you follow
  • One QA/RA contact for configuration and acceptance

What we decide together

  • Which regulators, feeds and document types to monitor
  • Update schedule and retention of superseded documents
  • Who may query, and where the audit exports are filed

What it is not

Not a document management system and not a replacement for your DMS or QMS. See the project page for what it does and what it does not.