Products

Privacy-first products, shipped.

Privacy-first principles, in practice.

01 · Live

Social Neuron

AI-powered social media management

End-to-end content creation platform that takes you from brand brief to scheduled posts across every major platform. Your brand inputs don't train anyone else's model.

Learn more Visit site

Key features

  • Zero-to-content in minutes from any brand URL
  • Multi-platform distribution: YouTube, TikTok, Instagram, Facebook, LinkedIn
  • Closed learning loop: performance from each post informs the next
  • Team workspaces with approval flows and audit trails
  • Privacy by design. DPIA completed, EU-hosted, no third-party trackers

Stack

React 19 · TypeScript · Supabase · Gemini AI · Remotion

02 · Live

TheVPNMatrix

Evidence-based VPN comparison

Independent VPN comparison built around a published evidence methodology. Every score traces back to a primary or independently verifiable source. The sources are linked from the page.

Learn more Visit site

Key features

  • 28-criteria evaluation · 3,469+ sourced evidence records
  • No affiliate-driven rankings: placement reflects evidence, not payouts
  • Methodology published: read the scoring system, reproduce the result
  • Quarterly audits: primary sources re-verified on a cadence
  • No third-party trackers: the comparison site doesn't surveil its readers

Stack

React · TypeScript · Tailwind CSS · Vercel

03 · Coming soon

Hearth

Local-first AI runtime

Run capable language models on your own hardware (laptop, desktop, on-prem). Hearth gives you context-aware AI that can read your files, notes, and tools without any data leaving the device. Privacy built into the architecture, attested at the boundary, not promised in a policy.

Request early access Private alpha · 2026
── Schematic · Hearth runtime
ON DEVICE THE NETWORK Files ~/DOCUMENTS Notes ~/NOTES Tools MCP Hearth ── LOCAL RUNTIME llama.cpp · MLX · Rust ↓ Answers, drafts DEVICE BOUNDARY × NO EXFIL by default Optional sync ── only if you ask
Solid line · in-process Dashed line · device boundary Coral · privacy invariant
hearth · ~/notes on-device
~ hearth chat
loaded 2,847 notes from ~/Documents/notes
models/Llama-3.1-8B-Instruct-Q5 ready on M2 GPU (12.4 GB)
no network calls authorised
› what did i decide about the Q3 budget?

reading · Q3 planning · Budget retro · Strategy 2026

Based on your notes from 14 March, you settled on £42k for Q3 with the contingency split 70/30 between hiring and tooling. You flagged this as conditional on the May review going well.

› draft a one-paragraph update for the team

saved to ~/notes/team-update-2026-05-22.md

~

Key features

  • On-device inference (CPU & GPU, Apple Silicon, NVIDIA)
  • Workspace memory that lives in a local SQLite vault
  • Native MCP support: connect tools without sending data out
  • Encrypted at rest, offline-first, sync only if you ask
  • Open methodology: every default explained, every model attested
  • Designed for individuals, regulated teams, and air-gapped deployments

Stack

Rust · llama.cpp · Tauri · MCP · SQLite

What is next

On the horizon.

We are exploring the next generation of privacy-first technology.

Local LLMs

Running language models on your own infrastructure, keeping sensitive data off third-party servers entirely.

On-device AI

AI inference that never leaves your device. Privacy built into the architecture, attested at the boundary, not promised in a policy.

Digital sovereignty

Tools that give individuals and organisations genuine control over their data, identity, and digital presence.

Want to work together?

Get in touch about partnerships, integrations, or research collaborations.

Get in touch