About
Frumu builds production AI systems and integrations for teams that want results they can run, monitor, and maintain—not a one-off demo. The focus is on clear interfaces, secure access patterns, and measurable acceptance criteria (quality, latency, cost).
Working style
- Integration-first: fit into existing product + data flows
- Clear acceptance criteria (cost/latency/quality)
- Security by design (PII handling, auditability)
- Docs + handoff for internal teams
Contracting
Frumu Ltd
Company Number: 16531590
71-75 Shelton Street, Covent Garden, London WC2H 9JQ, UK
What we've shipped (examples)
- RAG assistants with vector DB memory and tool calling
- Multi-agent editorial automation with review loops
- Automated research/extraction pipelines with structured outputs
- A production React Native mobile app for AI image/video workflows with a Django backend, WebSockets, subscriptions, and scalable deployment
Operational mindset
- Evals/regression tests where output quality matters
- Tracing/monitoring to debug retrieval and tool calls
- Cost/latency controls and safe fallbacks
- Deployment automation (Ansible) and scaling backends (Docker Swarm)
Engagement Types
Flexible engagement models to fit your project scope and timeline.
Project-based
Scoped AI integrations with defined deliverables and milestones.
£8,000 – £40,000+
depending on complexity
Retainer
Ongoing AI engineering capacity embedded with your team.
From £5,000/month
flexible commitment
Proof of Concept
Rapid feasibility testing before committing to a full build.
£2,500 – £5,000
fixed price, 1–2 weeks
Let’s scope your integration
If you’re integrating RAG, agent workflows, or multimodal pipelines into a production product, we can start with a short discovery and convert it into a concrete build plan.