A Shared Memory for Hermes and Claude Code
Hermes ships with a strong built-in memory system, but it lives inside Hermes. If you drive a second agent (in my case, Claude Code), the memory stays behind...
Browse by Topic
- Workflow 11
- AI-Agents 8
- Tutorial 7
- LocalLLM 6
- Strategy 5
- Benchmarks 3
- Case Study 3
- Guide 2
- Architecture 2
- OpenClaw 2
- Hermes 2
- LM Studio 1
- Ollama 1
- Cost 1
- Agent Workloads 1
- Patterns 1
- Diagnostics 1
- AgentTools 1
- Research 1
Workflow
Two Weeks of OpenClaw That Never Landed: The Day I Packed for Hermes
I packed my box for Hermes and it held almost nothing. Two weeks of OpenClaw, and the things worth carrying over fit in a single cardboard box. I spent the f...
Don’t Let Local LLMs Write Diffs: The L3c Pattern for Fat Skills
The moment I stopped letting gemma4:26b write patch_file calls, my skill stopped breaking. The fix wasn’t a bigger model — it was a three-layer responsibilit...
Hermes Agent Day One: Five Forks in the Road, Not Five Bugs
I ran Hermes Agent for a full day for the first time. The five places I tripped aren’t bugs — they’re forks in the road every adopter walks through on day on...
Hermes Agent in 5 Minutes: The One-Command Setup Guide
Yesterday I collected real user stories about Hermes Agent. Today I’m walking through the actual setup — one command, a few prompts, and you’re in.
I Actually Installed Hermes Agent. Here’s What Happened.
The setup guide I wrote yesterday was based on research. Today I ran the actual installer, connected it to Telegram, and tested whether the memory loop works...
My AI Stack in Spring 2026: Four Tools, Four Roles
I’ve spent the last three weeks wiring local LLMs into my daily work. Somewhere along the way, four distinct roles emerged — and the gaps between them told m...
The OpenClaw Configuration That Actually Works: Lessons from 6 Weeks of Daily Use
Most OpenClaw guides tell you what files to create. None tell you what to actually write in them. After 6 weeks of daily iteration, here’s the configuration ...
Hermes Agent Looks Interesting — So I Collected Real User Stories
I’ve been running OpenClaw daily for months. Hermes Agent keeps coming up. Instead of blindly switching, I went looking for what people who actually tried it...
I Let Claude Code Handle Everything I Was Too Scared to Touch
This looks risky. This looks like it’s only for engineers. That’s exactly what I thought — and exactly what AI is solving right now.
My Always-On AI Agent System: Telegram, Ollama, and an Obsidian Vault on a Mac Studio
I built a 6-agent AI system that runs 24/7 on my Mac Studio. Telegram for input, Ollama for inference, Obsidian for memory. Here’s the full architecture — ho...
AI-Agents
A Shared Memory for Hermes and Claude Code
Hermes ships with a strong built-in memory system, but it lives inside Hermes. If you drive a second agent (in my case, Claude Code), the memory stays behind...
Two Weeks of OpenClaw That Never Landed: The Day I Packed for Hermes
I packed my box for Hermes and it held almost nothing. Two weeks of OpenClaw, and the things worth carrying over fit in a single cardboard box. I spent the f...
Don’t Let Local LLMs Write Diffs: The L3c Pattern for Fat Skills
The moment I stopped letting gemma4:26b write patch_file calls, my skill stopped breaking. The fix wasn’t a bigger model — it was a three-layer responsibilit...
Hermes Agent Day One: Five Forks in the Road, Not Five Bugs
I ran Hermes Agent for a full day for the first time. The five places I tripped aren’t bugs — they’re forks in the road every adopter walks through on day on...
Hermes Agent in 5 Minutes: The One-Command Setup Guide
Yesterday I collected real user stories about Hermes Agent. Today I’m walking through the actual setup — one command, a few prompts, and you’re in.
I Actually Installed Hermes Agent. Here’s What Happened.
The setup guide I wrote yesterday was based on research. Today I ran the actual installer, connected it to Telegram, and tested whether the memory loop works...
My AI Stack in Spring 2026: Four Tools, Four Roles
I’ve spent the last three weeks wiring local LLMs into my daily work. Somewhere along the way, four distinct roles emerged — and the gaps between them told m...
Hermes Agent Looks Interesting — So I Collected Real User Stories
I’ve been running OpenClaw daily for months. Hermes Agent keeps coming up. Instead of blindly switching, I went looking for what people who actually tried it...
Tutorial
OpenClaw Auto-Reload: The Complete 6-Step Workspace Guide
Stop manually restarting gateways. Wire up launchd WatchPaths once, and every AGENTS.md edit auto-reloads both OpenClaw gateways in 30 seconds.
Running Llama 3.3 70B Locally: Hardware Requirements and Complete Setup Guide
Llama 3.3 70B is the most capable open-source model you can run at home — but it demands serious hardware. Here’s exactly what you need, what to expect, and ...
Complete Beginner’s Guide to Local LLMs: Everything You Need to Know in 2026
What are local LLMs, why would you run one, and how do you get started? A practical guide — primarily for Mac users — from zero to running your first AI mode...
Building Your Hybrid LLM Stack: Complete Implementation Guide
You understand the hybrid LLM concept. Now build it. This is the complete implementation guide — from installing your local models to deploying a team-ready ...
Ollama vs LM Studio 2026: Which Local LLM Tool Should You Choose?
A practical comparison of Ollama and LM Studio for running local LLMs. Features, performance, API compatibility, and which tool fits your workflow.
LM Studio Setup Guide 2026: How to Install and Run Local LLMs in 5 Minutes
A step-by-step LM Studio setup guide for Mac and Windows to run local LLMs. No cloud, no API keys, no monthly bills.
Best Local LLM Models for M2/M3/M4 Mac: Performance Benchmark 2026
Real benchmark data for running local LLMs on Apple Silicon. Token speeds, memory usage, and quality ratings for every Mac configuration from M2 Air to M4 Max.
LocalLLM
Before You Swap Your Local LLM Backend, Two Things to Check
If you’ve thought about switching the local LLM server on your Mac from Ollama to llama.cpp, there are two things that don’t show up in the obvious benchmark...
The Loop Tax Is Wall-Clock, Not Quality
I built a 4-branch falsification looking for the iteration ceiling on a local 27B-class agent. The quality cliff didn’t surface. The wall-clock divergence di...
Ollama and llama.cpp: Three Structural Differences on the Same Model
The community framing is simple: ‘Ollama is a ggml fork, llama.cpp is faster.’ I ran the same model blob through both runtimes on a Mac Studio M2 Max with th...
Is OpenClaw Actually Broken This April? Two Weeks of My Local Ollama Logs
OpenClaw has been brittle since late March. On my local Ollama setup I can now point at the exact line of code, the exact minute it went wrong, and the three...
The OpenClaw Configuration That Actually Works: Lessons from 6 Weeks of Daily Use
Most OpenClaw guides tell you what files to create. None tell you what to actually write in them. After 6 weeks of daily iteration, here’s the configuration ...
I Let Claude Code Handle Everything I Was Too Scared to Touch
This looks risky. This looks like it’s only for engineers. That’s exactly what I thought — and exactly what AI is solving right now.
Strategy
Stop Sending Everything to GPT-4: A 5-Factor Framework for Local vs Cloud LLMs
Stop sending everything to GPT-4. Five factors decide whether a task should run locally or hit a cloud API — here’s the framework to make that call in 30 sec...
LLM Cost Optimization: How to Reduce Your API Bills from $2,000 to $400/Month
A 5-person dev team was spending $2,000/month on LLM APIs. After applying these 7 techniques, they cut it to $400 — without losing output quality. Here’s exa...
Hybrid LLM Architecture: Save 50-70% on AI Costs with Smart Routing
Most teams route every AI task to GPT-4 or Claude. That’s like hiring a senior engineer to do data entry. Here’s the hybrid architecture that cuts API bills ...
GPT-4 vs Local Llama 3.3: Quality, Speed, and Cost Comparison 2026
GPT-4 costs $10-30 per million tokens. Llama 3.3 costs $0. But is the free option actually good enough? Here’s a side-by-side comparison across quality, spee...
Building Your Hybrid LLM Stack: Complete Implementation Guide
You understand the hybrid LLM concept. Now build it. This is the complete implementation guide — from installing your local models to deploying a team-ready ...
Benchmarks
Running Llama 3.3 70B Locally: Hardware Requirements and Complete Setup Guide
Llama 3.3 70B is the most capable open-source model you can run at home — but it demands serious hardware. Here’s exactly what you need, what to expect, and ...
GPT-4 vs Local Llama 3.3: Quality, Speed, and Cost Comparison 2026
GPT-4 costs $10-30 per million tokens. Llama 3.3 costs $0. But is the free option actually good enough? Here’s a side-by-side comparison across quality, spee...
Best Local LLM Models for M2/M3/M4 Mac: Performance Benchmark 2026
Real benchmark data for running local LLMs on Apple Silicon. Token speeds, memory usage, and quality ratings for every Mac configuration from M2 Air to M4 Max.
Case Study
My Always-On AI Agent System: Telegram, Ollama, and an Obsidian Vault on a Mac Studio
I built a 6-agent AI system that runs 24/7 on my Mac Studio. Telegram for input, Ollama for inference, Obsidian for memory. Here’s the full architecture — ho...
5 Models Tested, 2 Deleted: What Actually Works for Local AI Agents on M2 Max
Leaderboard scores don’t tell you which models work for AI agents. I tested 5 local models on my M2 Max for real agent tasks — orchestration, coding, researc...
I Run 3 Local Models and 1 Cloud API — Here’s How I Route Between Them
Theory says hybrid LLM routing saves money. I built a system that actually does it — 6 AI agents, 3 local models, 1 cloud API, running 24/7 on a Mac Studio. ...
Guide
Stop Sending Everything to GPT-4: A 5-Factor Framework for Local vs Cloud LLMs
Stop sending everything to GPT-4. Five factors decide whether a task should run locally or hit a cloud API — here’s the framework to make that call in 30 sec...
Complete Beginner’s Guide to Local LLMs: Everything You Need to Know in 2026
What are local LLMs, why would you run one, and how do you get started? A practical guide — primarily for Mac users — from zero to running your first AI mode...
Architecture
I Run 3 Local Models and 1 Cloud API — Here’s How I Route Between Them
Theory says hybrid LLM routing saves money. I built a system that actually does it — 6 AI agents, 3 local models, 1 cloud API, running 24/7 on a Mac Studio. ...
Hybrid LLM Architecture: Save 50-70% on AI Costs with Smart Routing
Most teams route every AI task to GPT-4 or Claude. That’s like hiring a senior engineer to do data entry. Here’s the hybrid architecture that cuts API bills ...
OpenClaw
Is OpenClaw Actually Broken This April? Two Weeks of My Local Ollama Logs
OpenClaw has been brittle since late March. On my local Ollama setup I can now point at the exact line of code, the exact minute it went wrong, and the three...
OpenClaw Auto-Reload: The Complete 6-Step Workspace Guide
Stop manually restarting gateways. Wire up launchd WatchPaths once, and every AGENTS.md edit auto-reloads both OpenClaw gateways in 30 seconds.
Hermes
Perplexity Cut Pro Deep Research to 20 a Month. My Hermes Agent Stack Runs Each Query for 22 Cents.
Perplexity cut Pro Deep Research to 20 queries a month in February 2026, a 900x downgrade. I rebuilt the loop on Hermes Agent, Exa, x_search, one local model...
X Search Used to Mean Scrapers, OAuth, or Paid Tiers. x_search Does It for $0.005.
xAI shipped x_search as part of their Agent Tools API. The price is $0.005 per query. Twenty dollars of credits covers roughly 4,000 searches. Here is what t...
LM Studio
LM Studio Setup Guide 2026: How to Install and Run Local LLMs in 5 Minutes
A step-by-step LM Studio setup guide for Mac and Windows to run local LLMs. No cloud, no API keys, no monthly bills.
Ollama
Ollama vs LM Studio 2026: Which Local LLM Tool Should You Choose?
A practical comparison of Ollama and LM Studio for running local LLMs. Features, performance, API compatibility, and which tool fits your workflow.
Cost
LLM Cost Optimization: How to Reduce Your API Bills from $2,000 to $400/Month
A 5-person dev team was spending $2,000/month on LLM APIs. After applying these 7 techniques, they cut it to $400 — without losing output quality. Here’s exa...
Agent Workloads
5 Models Tested, 2 Deleted: What Actually Works for Local AI Agents on M2 Max
Leaderboard scores don’t tell you which models work for AI agents. I tested 5 local models on my M2 Max for real agent tasks — orchestration, coding, researc...
Patterns
Don’t Let Local LLMs Write Diffs: The L3c Pattern for Fat Skills
The moment I stopped letting gemma4:26b write patch_file calls, my skill stopped breaking. The fix wasn’t a bigger model — it was a three-layer responsibilit...
Diagnostics
Is OpenClaw Actually Broken This April? Two Weeks of My Local Ollama Logs
OpenClaw has been brittle since late March. On my local Ollama setup I can now point at the exact line of code, the exact minute it went wrong, and the three...
AgentTools
X Search Used to Mean Scrapers, OAuth, or Paid Tiers. x_search Does It for $0.005.
xAI shipped x_search as part of their Agent Tools API. The price is $0.005 per query. Twenty dollars of credits covers roughly 4,000 searches. Here is what t...
Research
Perplexity Cut Pro Deep Research to 20 a Month. My Hermes Agent Stack Runs Each Query for 22 Cents.
Perplexity cut Pro Deep Research to 20 queries a month in February 2026, a 900x downgrade. I rebuilt the loop on Hermes Agent, Exa, x_search, one local model...