Skip to main content

OmniForge Blog

Field notes from the OmniForge team on building local AI for Mac.

When Your AI Agent's Memory Outgrew Your Ability to Trust It
Latest

When Your AI Agent's Memory Outgrew Your Ability to Trust It

We built an AI assistant with automatic memory, then hit the wall every agent system reaches: who decides what's worth remembering. Here's the two-layer architecture we settled on.

·9 min read

More articles

Letting an AI Agent Triage Production Alerts
·7 min read

Letting an AI Agent Triage Production Alerts

What we learned connecting Grafana to Discord to an AI agent and letting it handle incidents for 48 hours.

Green AI at the Edge: When Local LLMs Save Electricity and Water
·8 min read

Green AI at the Edge: When Local LLMs Save Electricity and Water

Right-sized inference beats the local-vs-cloud debate

OmniForge Desktop: Local-First Document Intelligence for macOS
·5 min read

OmniForge Desktop: Local-First Document Intelligence for macOS

Your documents and audio recordings in one private, searchable workspace on your Mac.

Your Local LLM Is Slow Because of Five Config Flags
·8 min read

Your Local LLM Is Slow Because of Five Config Flags

KV cache, context size, and batch size control more of the performance budget than most people realize

We Automated the Parts of Development Nobody Wants to Do
·5 min read

We Automated the Parts of Development Nobody Wants to Do

Automated code review, merge request creation, and standup summaries for GitLab, GitHub, and Slack.