AI features that ship to production, not just demos
I integrate OpenAI, Claude and Gemini into Vue and Nuxt products — assistants that know your data, agents that take actions, and generation pipelines with cost and quality under control.
What you get
- ◆
Assistants & agents
Chat UIs with tool calling that can query your data and trigger real actions.
- ◆
RAG on your content
Embeddings and retrieval over docs, products or user history with source citations.
- ◆
Streaming UX
Fast, streamed responses with good loading, error and retry states.
- ◆
Cost & safety controls
Model routing, caching, rate limits and evaluation to keep quality up and bills down.
How we work
- 01
Use-case scoping
Where AI adds real value — and where a rule is enough.
- 02
Prototype
A working slice on your data within days to validate quality.
- 03
Productionise
Streaming UI, guardrails, logging and evaluation.
- 04
Iterate
Prompt and model tuning based on real usage.
Selected work
FAQ
Which model should we use?
It depends on the task, latency and budget. I often route between models — a fast one for simple requests and a stronger one for complex reasoning.
Is our data safe?
I use providers’ no-training API tiers, keep keys server-side and minimise what is sent to the model.
Can you add AI to an existing app?
Yes — most projects are integrations into existing Vue/Nuxt products, not rewrites.
What does it cost to run?
I estimate token costs up front and add caching and routing so you know the monthly bill before launch.