Personal AI Infrastructure

Archer AI

Self-hosted intelligence platform. Conversational AI, price tracking, systems engineering, and home automation — powered by local GPU inference.

7
Services
10+
AI Models
25+
SAGE Tools
RTX 4070
Local GPU
UK
Self-Hosted
Intelligence Services

Hire the studio

Interactive experiences, AI-inside products, price intelligence, and systems engineering — built on the same NEXUS platform that powers these demos.

Custom intelligence for your project

See live demos, understand the process, then start with the Experience Compass or a direct conversation. UK-based, self-hosted, show-don't-tell.

Interactive demosAI integrationPrice trackingSysML
View services & demos →
Digital Experiences

Show, Don't Tell

Interactive pieces hosted on NEXUS — constraints in, meaning-driven experiences out. The infrastructure is the studio; these are what it makes.

The Naming Orb

A true story: two people, separate theme choices, one computed palette. Names re-rank where themes overlap; the orb's colours weave from what they share. Read-only snapshot — drag to orbit, tap a name.

3D canvasComputed paletteZero capture
Open the demo →

Experience Compass

Ten questions, one structured brief. Work out what kind of digital experience you actually need — landing page, tool, playable piece — before anyone quotes you a website.

Client intakeArchetype matchCopy / download
Try the compass →

The Boundary

Three constraints — Wedding, Money, Show-off — one triangle. Drag the puck; the geometry morphs and eight experience concepts illuminate by proximity. The Archer method, made explorable.

Ternary spaceLive morphZero capture
Explore the boundary →

Raw Geometry

The engine underneath: hand-rolled 3D morphs from seed head to icosahedron to torus knot. Organic drift, golden-angle growth, zero libraries. The whoa before the story.

WebGL-freeMorphing formsInteractive
See the geometry →
Platform

Intelligent by Design

A suite of microservices orchestrated by SAGE — a conversational AI that routes between local GPU inference and cloud models, with persistent memory and tool execution.

S

SAGE — Conversational AI

Multi-model interface with long-term memory, 25+ tools, and an OpenAI-compatible API. Routes between Ollama and cloud models based on task complexity.

Multi-turn chatTool callingMemorySSE streaming
£

Price Intelligence

UK retail price tracking with composite scoring, automated scraping, and deal detection. Monitors eBay, Scan, CCL with price-to-performance analysis.

UK retailerseBayDeal alertsGBP
M

Systems Engineering

SysML v2 modeller with text-canonical models, Lark parser, and Mermaid diagram rendering. Conversational model authoring via SAGE.

SysML v2MermaidLLM-first
H

Home Automation

Google Home ecosystem integration via Home Assistant. List devices, check state, and control services through natural language.

Google HomeHome AssistantVoice
Architecture

Privacy-First Infrastructure

Self-hosted on a single server with local GPU inference. All services bind to localhost — external access gated by Cloudflare Tunnel with zero-trust authentication.

Backend
Python / FastAPI
7 microservices
Inference
Ollama + Cloud
RTX 4070 12GB VRAM
Data
SQLite WAL
Per-service databases
Security
Zero Trust
CF Access + Tailscale
Development Log

Recent Updates

Latest from the NEXUS roadmap.