// 34 posts · no recycled content

ALL THE
REAL
STUFF.

MCP architecture, multi-agent orchestration, local inference, agent memory. Built in production. Described accurately.

MCP ArchitectureNEW

The Right Way to Build Multi-Tool MCP Servers (And the Three Most Common Mistakes)

After building 25+ MCP servers, here are the patterns that hold up in production and the shortcuts that blow up in your face when an agent starts hammering the endpoint. We cover tool schema design, error surface management, and why your retry logic is probably wrong.

18 min read March 2026 MCP · Production
// In This Post
  • Tool schema design patterns
  • Error surface management
  • Retry logic for agents
  • Rate limiting strategies
  • Auth without overhead
Read Post →
Local LLM
Running Qwen-Coder-30B on Dual RTX 3060s: Setup, Benchmarks, Limits
Real numbers, not marketing copy. What it can and can't do on consumer hardware with split-GPU inference.
Feb 2026
12 min read
→
Agent Memory
7 Types of Agent Memory (And When to Use Each One)
Episodic, semantic, procedural — not just theory. Built all 7, here's what each one actually changes at runtime.
Feb 2026
9 min read
→
Orchestration
Fan-Out/Fan-In: Running 10 Agents in Parallel with ThreadPoolExecutor
The pattern, the failure modes, and how to handle partial failure without blowing up the whole run.
Jan 2026
8 min read
→
Infrastructure
NATS JetStream vs. Redis Streams for Agent Message Queuing
Both work. Here's why we switched mid-project and what the tradeoffs actually are in a high-throughput agent loop.
Jan 2026
14 min read
→
Testing
Refactoring a 100x Coding Agent: 7 Modules, 34 Tests, Zero Regressions
How to safely refactor an agent system that's already in production without breaking the thing that's making money.
Dec 2025
11 min read
→
MCP
FastMCP Authentication Patterns: JWT, API Keys, and mTLS
Three auth strategies for MCP servers depending on your threat model and deployment context.
Dec 2025
10 min read
→
Local LLM
LM Studio vs Ollama vs llama.cpp: A Real Comparison for Agent Workloads
Not a tutorial. A decision guide based on running all three in the same agent loop with the same model.
Nov 2025
13 min read
→
🔒
27 more posts — members only Deep-dive architecture walkthroughs, full annotated code, and build logs from real client work.
Unlock for $97/yr →