SpaceXAI and Cursor

Grok and Cursor, built into the systems you already run

Deploy Grok, Cursor and Grok Bot within your business. Take advantage of advanced voice and image models. Supported by a real engineering partner.

Request a Quote

We reply within an hour during UK business hours.

What we usually find, and what we build instead

Four things we find in almost every company that has started using this stack before deciding how.

Without SpotDev

What we usually find

With SpotDev

What we build instead

Without SpotDev

Your team is already using it on consumer accounts

Consumer accounts train on what is typed into them by default. There is an opt-out. Nobody has clicked it.

With SpotDev

The right terms, chosen deliberately

Grok business and API terms state content is not used for training. On the coding side it is a setting, and we check which of your seats actually has it on.

Without SpotDev

Nobody has read which company you contract with

Grok business, Cursor and the consumer app are separate agreements with separate companies on different terms.

With SpotDev

Every agreement named before signature

Who you are contracting with, under which law, and what each one commits you to.

Without SpotDev

Data residency assumed rather than established

Buyers assume a UK option exists on the coding side. It does not, and the EU option that does exist is narrower than it sounds.

With SpotDev

Residency settled before the build

What can be processed where, what constraining it costs, and which features stop working when you do.

Without SpotDev

A model retires and the workload quietly changes

Retired model names redirect to a replacement and bill at the replacement's rate. Nothing breaks loudly.

With SpotDev

Pinned versions and a migration path

Model choice recorded, deprecations tracked, and a rebuild planned before a retirement date arrives.

What we build on this stack

The model is the smallest part of the work. Most of it is integration, permissions, grounding and proof, and that work is the same whichever vendor you land on.

Grok API integration

Agents and services built on the Grok API and connected to your CRM, finance system and document store under scoped access.

Cursor rollout for engineering teams

Single sign-on, provisioning, model and repository access policy, audit logging, and the review practice that stops output volume becoming review debt.

Voice agents wired to your CRM

Real-time voice built on the Voice APIs, with the transcript, the intent and the follow-up landing on the right record rather than creating a duplicate.

Grok Bot governance

Bots act on a cloud machine with a browser and a terminal. We define what they may touch and what needs approval. Team access is still rolling out, so we confirm you can buy it before scoping.

Data handling and residency

Retention settings, zero-retention where it earns its cost, transfer mechanism, and a written record of what leaves your estate.

Enablement and measurement

The people using it are trained on the changed workflow, and we capture a baseline before the build so the review compares like with like.

What to settle before company data goes near it

  1. 01

    Which agreements you are actually signing

    Grok business, Cursor and the consumer app are three separate contracts with separate companies. The vendor's own paperwork is mid-migration, and its enterprise terms and data processing addendum name different entities.

  2. 02

    Which terms your staff are actually on

    Business terms state content is not used to train models. Consumer accounts train by default unless somebody changes a setting, and UK users keep an opt-out most of them do not know exists.

  3. 03

    Where the data is processed

    Cursor documents US-only residency today, with EU and Iceland inference-only on request and no UK option. Constraining it carries a stated price uplift and a lead time.

  4. 04

    Whether zero retention is worth its cost

    It is not a free toggle. SpaceXAI's own docs say zero retention disables features including the stateful Responses API, Files and Collections, and the Batch API, and that most customers should not turn it on. Decide it deliberately.

  5. 05

    Whose account the bot runs on

    Grok Bot signs in with a Cursor account, and its training opt-out and retention follow Cursor's settings and terms rather than the SpaceXAI agreement you reviewed.

  6. 06

    What happens at a retirement date

    Model retirements redirect old names to a current model and bill at the new rate. A workload can change underneath you without an error.

Where the approval line sits

Authority is granted per action, not per agent, and permission to draft is never permission to send. These hold whichever vendor you end up on.

Runs unattended

Internal research, drafting, queue preparation and flagging. Nothing a customer sees and nothing that writes to a system of record.

Needs sign-off

Every external message, and any write to a record that carries money, a commitment or a contractual term.

Hard-blocked

Bulk or destructive changes, credential handling, payment actions, and anything outside the agreed source list.

Computer use is its own case

An agent driving a browser and a terminal can reach anything the session can. Scope the machine, not just the task.

Shared machines leak

Where bots share a cloud computer, files and logins can outlive the bot. Deletion of an agent is not deletion of its traces.

Everything is logged

Every action, approval and rejection recorded, so an incident can be reconstructed from evidence rather than recollection.

Fixed-price AI implementation packages

The same ladder applies whichever platform the work lands on. Model and platform usage is billed to you by the vendor and sits outside this fee.
  • AI Foundations

    £8,000

    Get AI deployed properly on the right platform for you, and your team actually using it.

    • Platform and model selection on your evidence
    • Workspace, access control and retention configured
    • Data handling and transfer position documented
    • Role-based training
    • Pre-build baseline for later comparison
  • Custom Agents

    £20,000

    A custom agent designed, built and deployed to production, integrated with your systems and governed properly.

    • Everything in AI Foundations
    • System and data source integration
    • Action-risk ladder written per action
    • Logging, evaluation and monitoring
    • Production handover and 60-day support
  • AI Transformation

    £45,000

    Multiple agents, an organisation-wide rollout and six months of optimisation support included.

    • Everything in Custom Agents
    • Multiple agents across departments
    • Rollout, enablement and adoption measurement
    • Deprecation tracking and migration planning
    • Six months of optimisation

Implementation expertise that drives adoption

An in-house engineering team that has delivered more than 300 technology projects. Work with engineers who have built this before.

300+HubSpot implementations, firm-wide
  • In-house engineering team.

    Agents and integrations designed, tested and shipped by our own engineers. Nothing is subcontracted.

  • Accredited where it is verifiable.

    HubSpot Diamond Solutions Partner, Custom Integration and Onboarding accredited, Cyber Essentials Plus, and 5.0 across 38 reviews on the HubSpot Solutions Directory.

  • No allegiance to this vendor.

    We hold no partner status with SpaceXAI, Cursor or Anysphere. If the answer is that you should not build on this, we say so.

AI with security in mind

  • Terms read before you sign
  • Cyber Essentials Plus
  • Approval boundary per action
  • In-house engineers

Common questions

Are you a SpaceXAI or Cursor partner?

No. We hold no partner status, accreditation or tier with SpaceXAI, Cursor or Anysphere, and we would rather say so plainly than let the page imply otherwise. What we do hold is HubSpot Diamond Solutions Partner, Custom Integration accreditation, Cyber Essentials Plus, an OpenAI Select Partner enrolment and certification in the Claude Partner Network, none of which confer anything on this stack. What we bring here is engineering, not a vendor relationship.

Can we buy licences through you?

No, and nobody else can either. Cursor's own pricing page states that subscriptions are sold only through cursor.com and that it does not authorise resellers or third-party sellers. You buy the licence direct from the vendor, and we do the engineering around it.

Can a UK company actually use Grok?

Yes. There is no UK or EU exclusion in the business terms, and the vendor's data processing addendum defines UK GDPR and applies the ICO's international data transfer addendum for UK transfers. One restriction we did find: custom voice cloning is documented as available only in the United States, with the exception of Illinois. If you were planning a cloned brand voice, that is not available to you here today.

Is our data used to train their models?

It depends which product, and this is the trap. Grok business and API terms state that content is not used to train models. Cursor is the other way round: with Privacy Mode off it may use code and prompts to train, and Privacy Mode is documented as on by default only for Enterprise teams, not for the Teams plan most mid-market buyers land on. Consumer Grok also trains by default, with an opt-out that UK and EU users keep even when signed out. Three products, three defaults.

Can we keep the data in the UK?

Not on the coding side today. Cursor documents US-only data residency, with EU and Iceland inference-only on request, no UK option, a stated price uplift for constrained residency, and a manual enablement lead time. If UK residency is a hard requirement, we will tell you that before you commit rather than after.

Which of the four should we start with?

Usually whichever has a cost you can name. Cursor is the easiest to justify because developer time is measurable and the rollout is bounded. Voice earns its place where call volume is real. Bots need the most governance work, so they are rarely the right first move.

What does it cost?

From £8,000 for a foundations engagement, £20,000 for a custom agent in production and £45,000 for a multi-agent rollout. Vendor platform and usage costs sit outside that and are billed to you by the vendor. Both vendors price in US dollars, so those figures move with the exchange rate.

Start with a diagnostic.

A short, fixed-scope assessment of whether this stack suits your obligations and your data, and where it would genuinely change how the business runs. No obligation, and sometimes the answer is that you should use what you already pay for.