Book a Health Check

OTTRA becomes an Anthropic partner

What the partnership changes for teams putting Claude to work inside a secure GitLab environment.

Illustrative image of a person at a laptop marked with the OTTRA logo and Anthropic and Claude stickers.

The short version

We are excited to announce that OTTRA is an Anthropic partner. The platform your code lives in and the AI that works on it are set up, secured and operated by one team, with the same boundaries and the same record of who decides what.

On this page

What the partnership is

OTTRA is an Anthropic partner, as of 13 November 2026. Anthropic runs the Claude Partner Network for companies that help organisations with AI strategy, consulting and Claude implementation.

We already build, secure and operate GitLab for engineering teams. Bringing Claude into that work is the same job: agreed access, clear boundaries and a named person who decides. The partnership puts us inside Anthropic’s supported routes for doing it, rather than improvising around them.

OTTRA is a GitLab Select and Professional Services partner, an Anthropic partner and a UKAI member.

Talk to us

Why this matters for your teams

Most organisations adopting Claude end up with a split supply chain. One supplier sells model access. Another runs the engineering platform. The join between them belongs to nobody.

That join is where the difficult questions live. Which repositories may Claude read? What leaves your network, and under what terms? Who approves a change it drafts? What can you show a reviewer in six months?

We hold both ends. One team sets up the platform, the access route and the review practice, and answers for all three. When a security lead asks where the code goes, there is one place to ask.

The hard part is the boundaries

Buying Claude is straightforward. Deciding what it may see is engineering work, and it is the part teams usually skip.

  • Permitted context. The repositories, branches and data a task may use, agreed before the task starts, with secrets and production data excluded.
  • A named reviewer. Claude drafts. A person reads the diff, checks the evidence and decides whether to accept it.
  • A record. The platform records what changed and who approved it, beside the work, in the same system as everything else.

This is what Claude Enablement sets up, using a real task from your own backlog rather than a sample exercise. Teams leave with a setup record, review guidance and a plan for the next task.

Where Claude runs is a separate decision

Access route and data location are their own question, and partnership does not settle it for you. GitLab hosting does not determine Claude’s inference location. Check the provider, model, endpoint, routing, retention and terms for each service separately.

We work through that choice with you on Cloud Choice™, alongside where GitLab itself runs.

What you can ask us for

  • Claude Enablement: get your engineers using Claude and Claude Code on an agreed task, with access, boundaries and review set up properly.
  • Cloud Choice™: we operate GitLab, Claude access, or both, where you need them, with data, access and ownership kept clear.
  • Health Check: a review of your platform, delivery and AI readiness before you commit to a direction. A paid engagement, scoped with you first.

What it does not change

A partnership is a supplier relationship, not a result. We would rather say plainly what it is not.

  • It is not a guarantee of an outcome. Scope, timing and cost are agreed before work starts.
  • Your engineers still decide. Nothing reaches your main branch because a model suggests it.
  • Cloud accounts, billing, subscriptions and usage charges remain yours and stay separate from our service.

If you are weighing up where AI fits in your delivery system, start with a bounded task you can actually check.

Primary sources

Programme scope and product terms can change. Checked on 13 November 2026.

Partnerships & assurance

Scroll to see allUse the left and right arrow keys to scroll.