Guides

Hands-on walkthroughs that go deeper than a single tool. Each one answers a real question a working developer has, with worked numbers and concrete steps, and links to the tools here that do the work.

• JSON JSON vs JSONL (JSON Lines): When to Use Each, With Real Examples JSON vs JSONL (JSON Lines/NDJSON): why fine-tuning, batch jobs, and logs use one object per line, with real examples and how it relates to NDJSON. July 16, 2026
• JSON JSON Schema Validation Explained: How to Catch Bad Data Before It Breaks Your API JSON Schema validation done right: draft 2020-12 vs draft-07, JSON Pointer error paths, and why `format` often isn't checked. With a hands-on validator. July 13, 2026
• Regex Catastrophic Backtracking and ReDoS: How to Tell If a Regex Will Hang (With Measured Timings) Why (a+)+$ hangs on a 30-character string: the two structural red flags for ReDoS, measured backtracking timings, and the de-nesting fix that scales to 100k chars. July 13, 2026
• LLM Counting Tokens Before You Call the API: tiktoken vs Anthropic vs Gemini, and Why the Numbers Differ Why the same text tokenizes differently across OpenAI, Claude, and Gemini, and how to count tokens accurately before you call any LLM API. July 9, 2026
• LLM Tokens Per Second: What Speed to Expect From Local LLMs on Your Hardware Estimate local LLM tokens-per-second from memory bandwidth and model size. A durable formula for whether a model is fast enough to actually use. July 6, 2026
• LLM KV Cache and Context Length: Why a Bigger Context Window Eats Your VRAM The KV cache grows linearly with context and concurrency. The GQA-aware formula, worked serving math, and why it caps throughput before weights do. July 2, 2026
• LLM GGUF Quantization Explained: Q4_K_M vs Q5_K_M vs Q8_0 — Which to Pick Decode GGUF quant suffixes (Q4_K_M, Q5_K_M, Q8_0) and pick the right one by VRAM headroom, not guesswork. Sizing driven by a live calculator. June 29, 2026
• LLM Which LLM Can You Actually Run on Your GPU? A VRAM-Tier Fit Guide Start from the GPU you own and find which LLMs fit, fit with offload, or won't fit. A VRAM-tier reverse lookup with real Q4 weight math. June 25, 2026
• LLM How much VRAM do you need to run Llama 3 or Gemma locally? The real math behind local LLM memory: weights, KV cache, and overhead, with worked numbers for Llama 3 8B and Gemma 2 9B. June 22, 2026
• LLM Self-Hosting a Local LLM vs Paying for an API: Where's the Break-Even? When does self-hosting an LLM beat a pay-per-token API? The real break-even math on GPU cost, throughput, and volume, with a calculator that does it for you. June 22, 2026