On-premise: your hardware, your network
A deployment on servers you own, inside your network, with open-weight models served locally and an air-gapped option where the data may not leave the building.
When it fits
- Data that may not leave the network, by contract or policy
- High, steady volume where owned hardware pays off
- A team that already runs its own servers
- Air-gapped requirements
The reference architecture
The named parts. Sizes and figures come from the assessment, not from this page.
Hardware
Workstation-class for a small team; server-class with one or more GPUs for volume. Sized in the assessment.
Model serving
Ollama or vLLM serving open-weight models pinned to a version you approve.
Vector store and database
Qdrant or Postgres with pgvector on the same host or a neighbour.
Application services
A containerised FastAPI service and a small web front end, reachable only inside your network.
Updates
Pulled on a cadence you set; in an air-gapped environment, carried in by hand.
Who holds what
Keys and data
Never leave your network. No telemetry, no call home.
Uptime
Yours. Runbooks cover restart, backup and restore; optional run means we do it on a schedule.
Models
Pinned versions; an upgrade is a decision, not an event.
Backups
Local snapshots to a location you choose, with a restore tested at handover.
What you give up
- Frontier-model quality is not available; open-weight models are chosen by measurement on your tasks and the gap is stated.
- Capital cost up front and someone in-house owning the hardware.
- Updates are your cadence; an air-gapped environment updates by hand.
When another option is the better answer
Cloud
You own it, we build it
Choose cloud when you have no infrastructure team, when volume is variable, when you need a provider's frontier model through a private endpoint, or when a named cloud region satisfies your residency obligation.
CloudManaged agents
We run it
Choose a managed agent when the report is what you want and nobody on your side wants to own the pipeline. We host it, monitor it and send the output on schedule.
Managed agentsWhat this is not
The TurnkeyGTM platform runs as a cloud service we operate. We do not offer an on-premise or private-hosted edition of the platform today. On-premise and cloud deployments are designed and built on open-weight models and self-hostable components — containerised services, a vector store such as Qdrant or Postgres, and a model server such as Ollama or vLLM — and handed over with documentation. Managed agents run on infrastructure we operate, under a data-handling rule set agreed before the first run. Capacity, latency and cost figures are given per engagement, never on this page.
Questions we get asked
The capabilities behind it
Custom hosted and on-premise AI
AI deployments designed for where your data has to live — a dedicated hosted environment or your own hardware — on open-weight models and self-hostable components.
See the deliverablesModel selection and cost optimization
Which model, at which size, for which step — chosen by measured quality and cost on your own tasks rather than by the leaderboard.
See the deliverablesPII protection and data handling
Redaction, minimisation and handling rules for personal data in transcripts, tickets and documents before any model sees them, with residency where your obligations require it.
See the deliverablesTell us where your data has to live.
The assessment is short and fixed price. It ends with a written recommendation and a reference architecture.