Blog
Deep dives on local-first AI agents
How to Run Qwen 3.8 27B Locally
Run Qwen 3.8 27B as a tool-using agent on your own machine: memory math for 24 GB cards, grammar-constrained tool calls, and two measured agent runs.
Read article →Atomic Agent vs Hermes vs OpenClaw: Local Agents Compared
Atomic Agent, Hermes, and OpenClaw compared: philosophy, local-first stance, licensing, and how each held up in the GAIA L1 head-to-head.
Read article →GBNF Grammar-Constrained Tool Calling, Explained
GBNF grammars make it structurally impossible for a local model to emit an invalid tool call. Here's the mechanism that turns small models into reliable agents.
Read article →How to Run an AI Agent Offline (No API Key, No Token Bill)
Run a capable AI agent fully offline via llama.cpp: no network, no API key, no token bill. Step by step, and it still drives your files, shell, and git.
Read article →Local AI Agent vs Cloud AI Agent: The Real Trade-offs
Local vs cloud AI agents compared on privacy, cost per token, latency, offline use, and model ceiling. An honest side-by-side table, not a sales pitch.
Read article →What Is a Local AI Agent? A Plain-English Definition
A local AI agent runs its model, reasoning loop, and tool execution entirely on your own machine: no cloud round-trips, no API keys, no per-token bill.
Read article →