On-premises · your servers · your LLM key

On-premises AI
software delivery

For banks, insurers, healthcare, government and every company whose code cannot live in someone else's cloud: Squadship deploys its full AI delivery squad on your servers, inside your network, on your own LLM API key. Not an enterprise tier of a SaaS product — this is the product.

Request a demo What is Squadship? →

The data flow, disclosed

Security reviews shouldn't need archaeology. Here is exactly what stays, what leaves, and on whose terms.

Stays on your infrastructure

  • Every repository, document, board and decision log
  • All generated work — PRDs, designs, mockups, code, test results
  • The vector indexes built from your codebase
  • User accounts, sessions and audit history

Leaves your network — on your terms

  • Task-relevant excerpts to the LLM, under your own API key and provider agreement
  • A license heartbeat to the vendor: status and usage counters, field list disclosed in the agreement — never code or documents
  • Nothing else. No vendor cloud in the data path

Enterprise controls, in writing

The boring paperwork that real procurement asks for — answered before you ask.

Control

  • Bring your own model provider — you choose the model, the terms, the spend
  • Per-project token metering: see exactly where LLM spend goes
  • Role-based access, org sign-in policy, SSO with MFA
  • Humans approve every gate; nothing merges or ships unsigned

Assurance

  • License, support schedule and telemetry field list — all in writing
  • Graceful, disclosed expiry: data readable and exportable at every stage
  • Decision logs, signed baselines, requirement-to-release traceability
  • Audit-ready activity records, written as the work happens

Common questions

What does on-premises mean for Squadship exactly?
The whole platform — agents, documents, boards, generated code, vector indexes — runs on servers you control, inside your network. There is no vendor cloud in the data path. On-premises isn't an enterprise tier of a SaaS product; it is the product.
What about the AI model — doesn't that need the cloud?
The model runs under your own LLM API key, with your own provider relationship and data terms. Only task-relevant excerpts are sent per request, and token usage is metered per project so you always know where the spend goes. If your provider offers a private or regional endpoint, you can use it.
What data does the vendor receive?
A small license heartbeat: license status and usage counters, with the exact field list disclosed in the license agreement. Never your code, documents, or generated work.
What happens to our data if we stop using Squadship?
License expiry is graceful and disclosed up front: capabilities step down in stages, and your data stays readable and exportable at every stage. Your code was in your own Git repository all along, so there is nothing to migrate out.

Bring the squad inside your firewall

A live demo, then a 15-day pilot on your infrastructure with success criteria agreed up front — and your security team welcome in the room.

Request a demo