Pulse Blog
Notes from a local-first AI platform
Engineering writeups, agent design patterns, and lessons learned building Pulse. Honest, technical, no hype.
14 ways Pulse prevents LLM hallucination
Concrete walkthrough of the grounded auto-fires Vex runs before the LLM speaks. Real examples for each.
9 min read
How Vex's autonomous loops work
Scheduled goals: daily brief at 07:00, hourly summaries, autonomous chat notifications. The agent works while you sleep.
7 min read
Local-first AI, explained without the buzzwords
Why running models on your own hardware beats sending every prompt to a US cloud — and where it doesn't.
6 min read
How Vex's 25 tools work together
The agent loop, tool dispatch, and the 7 grounded auto-fires that prevent hallucination.
8 min read
Using n8n as your AI tool layer
Webhook workflows + REST PUT cache invalidation. Why we picked it over building from scratch.
7 min read
Want to be notified when articles drop?
Sign up for a Pulse account — we'll ping you in your workspace when a new piece publishes.
Create your workspace