AI assistants, automations and the controls around them: approvals, logs and a way back. Built for teams that run shops, catalogs and back-offices.
Available for UK, EU and Asian companies. Remote delivery, CET to ICT.
OLL4 was accepted into Anthropic’s Claude for Startups program. OLL4 is enrolled in the OpenAI Partner Network.
OLL4 is a member of NVIDIA Inception.
OLL4 (INTRATRADE SINGLE MEMBER P.C.), since October 2026. Membership in a program is not an endorsement of OLL4 or SpaceApp by Anthropic, OpenAI or NVIDIA.
© 2025 NVIDIA, the NVIDIA logo, and NVIDIA Inception are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and other countries.
Scoped, with handoff rules and review points.
Intake, routing, drafting, approvals, exceptions.
Catalog, supplier and marketplace flows, any shop platform.
One view of health, queues and drift, with runbooks.
Releases you can reverse. Access you can explain.
Agent chat, live terminals, a managed browser and the resulting files, together in rooms that persist. Supervise the work, take over, come back later.
npx --yes run-spaceapp@latest installOur own control surface for live operations. Signals, decisions and follow-through in one place. Diagnostics are read-only; anything sensitive waits for a person.
Not for sale. Shown as a sanitized walkthrough during discovery.
A public AI endpoint built to be safe. Same-origin only, rate limits, a bot check when traffic looks automated, short sessions, no actions anywhere. It prepares a summary; a person reads it.
Map one workflow: what happens today, what fails, what to measure.
You keep: a written system boundary.
The smallest end-to-end slice on real inputs, with logs from the first run.
You keep: a working pilot.
Approvals, permissions, monitoring and a rollback that has been tested.
You keep: a runbook.
Measure real behaviour, remove the next bottleneck, hand over or continue.
You keep: a system your team runs.
Our entry for the Future Health Digital Innovation Challenge, statement “Closing the Screening-to-Care Loop”. Fixed, cited rules find overdue HbA1c follow-ups in diabetes care; the app drafts the message, and a clinician approves each action before anything leaves it.
Health, developer-tooling and agent-safety challenges. We name a challenge here only after our entry is registered and its rules allow us to say so.
Only what is built and running. We claim no prize, ranking or endorsement before the organisers publish results. A health prototype is not clinical advice.
Remote
APIs, interfaces and agent tooling for SpaceApp and client systems. Postgres, tests from the first commit.
Remote
Linux, Docker, CI/CD. Releases with a tested rollback, monitoring and access you can explain.
Remote
AI assistants with tools, evaluations and approval points. Guardrails and cost limits.
Remote
Catalog, supplier and marketplace flows across shop platforms and ERPs. APIs, queues, retries.
Remote
End-to-end checks in real browsers, regression suites, release gates.
Which workflow creates the most manual effort, delay or risk right now?