GLM-5.3-Flash is Z.ai's newest open model, released in August 2026 under a plain MIT licence. It is also big. With 321 billion parameters, of which 18 billion are active per token, it needs roughly 160GB for a usable 4-bit build, so most 128GB machines can only run it at 1 or 2 bits. This guide covers what it needs, which machine runs which build, the one Ollama surprise to … [Read more...] about How to Run GLM-5.3-Flash Locally: 321B, MIT Licence, and the Memory It Takes
Tutorial
How to Run Qwen3.8-Flash-Next Locally: Memory, Speed and the Licence Catch
Qwen3.8-Flash-Next is the most interesting open model of the autumn for anyone with a big-memory machine. It has 180 billion parameters in total, but only 6 billion are active for each token, so it can generate text quickly on hardware that would crawl through a dense model of the same size. Qwen calls it an experimental preview of the architecture behind Qwen4. This guide … [Read more...] about How to Run Qwen3.8-Flash-Next Locally: Memory, Speed and the Licence Catch
Stable Diffusion LoRA: Matching the Base Model, Weights and Stacking
Quick answer A LoRA only works on the base model it was trained against. An SDXL LoRA will not load onto SD 1.5 or Flux. This is the single most common reason a LoRA "does nothing". You never have to guess which base it needs. Every LoRA on Hugging Face declares a base_model field in its metadata — check it before you download. Strength is a dial, not a switch. A scale … [Read more...] about Stable Diffusion LoRA: Matching the Base Model, Weights and Stacking
ComfyUI Templates in 2026: 660 Built In, and Half Need the Cloud
Quick answer ComfyUI ships 660 workflow templates in the version current at the time of writing — far more than the handful most guides show you. 🔴 368 of them (56%) are cloud API templates. They send your prompt, and often your image, to Google, Grok, MiniMax, Recraft or similar. On a tool people install specifically to run things on their own machine, the majority of … [Read more...] about ComfyUI Templates in 2026: 660 Built In, and Half Need the Cloud
Best Vector Database for Local RAG in 2026: What It Costs in RAM
Quick answer For a personal RAG corpus the database is not the expensive part. Fifty thousand chunks embedded at 768 dimensions need roughly 188 MiB of RAM once index and overhead are counted — less than the embedding model that produced them. Four different kinds of software get called “a vector database”. Choosing the wrong type costs you far more than … [Read more...] about Best Vector Database for Local RAG in 2026: What It Costs in RAM
Stable Diffusion Prompts: What the Documentation Actually Says
Most Stable Diffusion prompt advice is folklore: lists of magic words, copied "quality tags", and parentheses stacked until something changes. Some of it works by accident. This guide covers what the documentation and the model configuration files actually say — including a hard limit on prompt length that nobody mentions, and which is the real reason your carefully written … [Read more...] about Stable Diffusion Prompts: What the Documentation Actually Says
Best Local Text-to-Speech Models in 2026: Sizes, Licences and What Runs Offline
Local text-to-speech has quietly become practical. A usable English voice is now a 63 MB file that runs on a CPU, and the best-sounding open model is 327 MB. The catch is not quality or hardware any more — it is licensing, and one of the most widely used models is non-commercial in a way that is easy to miss, because the licence it points at no longer exists on the web. Quick … [Read more...] about Best Local Text-to-Speech Models in 2026: Sizes, Licences and What Runs Offline
Five Ways a Real-World PDF Breaks pdftotext (And How to Read It Anyway)
Every tutorial on extracting tables from PDFs uses the same example: a clean, machine-generated invoice where pdftotext -layout returns tidy columns and everyone goes home happy. Then you point the same script at a document somebody actually made, in InDesign or Illustrator, for print — and it returns rows of blanks. Or worse, rows that look correct and aren't. I spent a few … [Read more...] about Five Ways a Real-World PDF Breaks pdftotext (And How to Read It Anyway)
Self-Hosting n8n with Local AI: Setup, Licence and What You May Not Do
n8n is the usual answer when someone wants to wire a local model into real automation — trigger on an email, pull a document, ask the model, write the result somewhere. Self-hosting it is genuinely easy. The part almost every tutorial skips is that n8n is not open source in the sense people assume, and its licence says something specific about what you may do with it. This … [Read more...] about Self-Hosting n8n with Local AI: Setup, Licence and What You May Not Do
Local AI Pair Programming in 2026: Which Tools Are Still Maintained
Search for an AI pair programmer that runs against your own local model and you will be pointed at the same tool almost every time. It has 49,000 stars, it is genuinely good, and it has not shipped a release in more than a year. Nothing on its page says so. This guide checks which terminal and editor AI assistants are actually still being developed, using repository activity … [Read more...] about Local AI Pair Programming in 2026: Which Tools Are Still Maintained