Wiring a Local LLM into n8n Without a Single API Key
Connecting a self-hosted OpenAI-compatible model to n8n: container networking, the request shape that actually works, timeouts, and the smoke test that saves an afternoon.
Read article →Blog
Notes on local AI, automation, and shipping open source. Subscribe via RSS.
How to Write System Prompts That Actually Work
Learn why structuring system prompts with XML tags improves LLM consistency, reduces hallucinations, and produces more reliable outputs.
Read article →Unlocking the Beast: Pushing AMD Strix Halo to 120GB Unified VRAM under Linux
Pushing AMD Ryzen AI Max+ 395 (Strix Halo) to 120GB unified GPU VRAM under Linux with kernel GTT tweaks, BIOS adjustments, and ROCm configuration.
Read article →Introducing zBudget: Self-Hosted Personal Finance, Reimagined
A self-hosted budgeting app built with React, FastAPI, and PostgreSQL — what it does, how it compares to Actual Budget and Firefly III, and when you should pick something else.
Read article →Six n8n Workflows You Can Import in a Minute
A walkthrough of the six n8n templates in our workflow library — what each one does, the reliability patterns baked into the JSON, and how to adapt them.
Read article →Why We Went Fully Open Source
Our journey and reasoning behind making all Zypher Systems projects transparent and community-driven.
Read article →