HN Showcase
Back to browse
GitHub Repository
3 stars

Own your AI's context and memories across every model and device

by kaspern·Mar 9, 2026·7 points·1 comment
Visit ProjectView on HN

AI Analysis

●●●BangerZero to OneBig BrainSolve My Problem

Portable AI memory across Claude, GPT, Gemini using your own Postgres knowledge graph.

Strengths
  • •Genuine multi-model portability—context actually follows you between vendors, a real pain today.
  • •Open-source MCP server with sensible architecture (pgvector + semantic search, $6/month VPS).
  • •Three months of daily use proves it works in practice, not just theory.
Weaknesses
  • •Requires manual setup (Supabase, VPS, TypingMind integration) — not one-click for non-technical users.
  • •Privacy pitch is strong but long-term maintenance and scaling of personal infra is user's problem.
Target Audience

Power users of multiple AI models who want portable context and privacy from vendor lock-in

Similar To

Anthropic's MCP Protocol (ecosystem standard) · Cursor's context awareness · Continue's codebase integration

Post Description

Hey HN,

I pay for ChatGPT, Claude, Cursor, and use Gemini through work. Four vendors, four separate conversation histories, four profiles of how I think. None of them talk to each other. Switch providers and you start over.

So I built a system where the memory is mine. I run a knowledge graph in Postgres (Supabase, free tier) with pgvector for semantic search. A small MCP server reads and writes to it. That server sits behind an MCP Gateway on a $6/month VPS, along with Brave Search and a GitHub server. TypingMind connects to the gateway as a BYOK client -- any model, any device, same brain.

https://github.com/kaspnilsson/digital-twin-playbook

When I switch from Claude to GPT-5 to Gemini, the new model picks up where the old one left off. After three months of daily use, the AI knows my project architectures, my preferences, my side projects. I never re-explain any of it.

The MCP server is MIT-licensed: https://github.com/kaspnilsson/mcp-memory-supabase The playbook walks through the full setup -- Supabase schema, VPS hardening, Caddy, systemd, the system prompt that makes tool routing work: https://github.com/kaspnilsson/digital-twin-playbook

What it costs me: ~$45/month ($6 VPS + ~$36 API compute via OpenRouter + $3 amortized TypingMind license). More than a $20 subscription. But the $20 price is subsidized, and my data stays on my server. What does not work well: TypingMind is a PWA, not a native app. Voice is rough. iOS kills background processes on long tool chains. MCP config does not sync across devices. You are your own SRE.

This is not a consumer product. If you want polish, use Claude.ai. If you want to own the context that makes your AI useful, this is how I did it!

Similar Projects

AI/ML●●●Banger

Sulcus Reactive AI Memory

Thermodynamic memory decay beats passive vector search—90% token reduction claimed.

Big BrainWizardry
mcdoolz
403mo ago
AI/ML●●Solid

Sekha – What if AI remembered 3 years of conversations, not 3 hours?

The core idea is simple and pragmatic: attach a persistent, SQLite-backed vector store to any model so conversations don't vanish after a single context window. The repo leans into portability (Rust, self-hosted, AGPL) and the UI shows sensible controls like conversation folders and a context-budget token slider — useful details that suggest this is built for real use rather than a demo. My worry: retrieval quality, scaling and access controls will be the real battleground, not the clean chat UI.

Niche GemSolve My Problem
sekha-ai
113mo ago