I've been working on a project called Maker, which is an attempt to rethink personal automation in the AI era.
Most automation tools (Zapier, n8n, Make) focus on connecting APIs.
Most AI tools focus on generating responses.
Maker focuses on something slightly different: operating your own automation workflows.
Key ideas:
1. Local-first desktop app (SQLite, no required cloud)
2. Fast-first reporting (no 60-second LLM blocking)
3. Granular permission model (per-bot, per-capability)
4. BYO LLM — you control cost and data egress
5. AGPL-licensed core (network use requires source disclosure)
The goal is to build a personal automation operating system — not a bot marketplace.
I'm particularly interested in feedback on:
- Local-first architecture choices
- AGPL vs BSL tradeoffs
- Whether "personal automation OS" is a useful framing
I've been working on a project called Maker, which is an attempt to rethink personal automation in the AI era. Most automation tools (Zapier, n8n, Make) focus on connecting APIs. Most AI tools focus on generating responses. Maker focuses on something slightly different: operating your own automation workflows. Key ideas: 1. Local-first desktop app (SQLite, no required cloud) 2. Fast-first reporting (no 60-second LLM blocking) 3. Granular permission model (per-bot, per-capability) 4. BYO LLM — you control cost and data egress 5. AGPL-licensed core (network use requires source disclosure) The goal is to build a personal automation operating system — not a bot marketplace. I'm particularly interested in feedback on: - Local-first architecture choices - AGPL vs BSL tradeoffs - Whether "personal automation OS" is a useful framing