All Activity

@ryan that is exactly the one I have been playing with today. The 14b version. Unfortunately, qwen 3 coder only comes in 30b and 40b, which is pushing it for me. I might try it again with Ollama. It suffered with LM Studio, but Ollama is meant to be better at dealing with local problems like not having a huge stack of GPUs lying around! The hard bit is the prompting. I was trying to get it to do a very simple page. Create a data set, then search it and spit out a single result. It was determined to ignore the search result and just loop the dataset. Grr! EDIT: By the way, when I installed, it failed to create the subdirectories in jobs. I have hand created them and copied the htaccess file from jobs into each one. Is that right? Anyway, here is the reason for the large box, and the view out of my studio window! Let me know when you start a podcast and need an editor!

Thankyou @pideluxe

@Joss Wow, that sounds like a nice hardware setup you've got there. And 128gb ram is like having a bag of gold right now, nice pirates booty. Have you tried Qwen 2.5 coder yet? Not a new model, but seemed pretty good when I tried it: https://ollama.com/library/qwen2.5-coder

Thank you, kind sir! (The Pirate bows!) I will have a play. I have been using qwen instruct but also have a few other models. I am a little limited because I only have an RTX 5080. Google's AI swore blind I could run a 31b model with a fair amount of offset (I have 128gb Ram), but I last saw the machine crawling out from under the desk, sweating and groaning, so I am playing with the 14b instead with 1.5b for autocomplete! I will get there. Honest! I still guess at all this stuff.

@Joss Yes, on the "Agents" tab, enter a new Agent with the settings (replacing the model ID with whichever model you are using): Model ID: llama3.1 API key: ollama Endpoint URL: http://127.0.0.1:11434/v1 Label: Ollama llama3.1 The API key can be any non-empty value, like ollama, because Ollama usually doesn’t require one. AgentTools requires the field to be populated. The endpoint can be either one of these: http://127.0.0.1:11434/v1 http://127.0.0.1:11434/v1/chat/completions The model ID must match an installed Ollama model name from ollama list. Note that the Engineer relies on tool/function calling. So the Ollama model needs decent tool-call support, or it may answer text but fail to use eval_php, site_info, etc. For AgentTools, a coder/instruct model with tool support is the safer choice. Btw, if the connection works but Engineer behaves oddly, try a stronger tool-capable local model before assuming the config is wrong. Local Ollama can be great, but AgentTools asks a lot of the model because it needs reliable structured tool calls.

Hi Ryan! Will this, or something, talk to Visual Studio Code via Continue and Ollama? (I have LLMs locally). I am still trying to get my head around what is possible for site development with AI and Processwire - not helped that I am so out of date now!!! Joss

Hi everyone! Just wanted to get a general feel on the documentation pages of the new website, is anyone else finding troublesome that the main sections of the documentation are now collapsed? Weren't these sections "less collapsed" or maybe I an hallucinating (i know it's in fashion) ? Wouldn't it just be better to have all of this expanded, in a big chunk of text? I really feel it makes it hard to use the documentation pages, wouldn't it just make more sense to have submenus under the API variables? Here's a quick mockup of what I think could be a more pragmatic layout:

Revisiting this topic again... I want to use vanilla CSS as much as possible moving forward, but I'm trying to find a "base" from which to start. This base would come with an opinionated way of organizing files, a reset, breakpoints and perhaps stub-out certain components and use all the latest CSS features. I can't seem to find anything like that and that's also pretty recent. Any suggestions?

👍Then use "Schlagwort / Schlagwörter".

Hmm, interesting. Thanks @pideluxe I actually have Tags reserved for a different part of the Module. So just to recap Tags = used for front end display like you probably already understand. Tagging images in an images field. Labels = similar purpose but used for Library organisation in the admin area

Actually, we use "Etiketten" just for physical labels. One could use "Schlagwort / Schlagwörter" or "Tag / Tags" for more tech-savy users.

Thank you in advance! Nothing is urgent, of course :)

I'll change that as soon as I've time.

BTW is this the best translation for Label and Labels? Google seems to suggest a few variations Label → Etikett Labels → Etiketten P

Hi If you are using 1.19.50 Beta, you may have experienced 2 translation issues. These are fixed and scheduled for the official 1.19.50 release, but if you're stuck, please DM me. German language CSV import fails with 'Undefined array key' warnings German translation: glossary fix — Label → Etikett, Labels → Etiketten Vielen Dank fürs Lesen, i believe 🙂

ok... it seems that this is not the problem, and we've moved on.

I am trying to learn php and "proper" dev practices (terminal, git, staging). My background is in no-code apps, but I have a solid knowledge of HTML and CSS. I started building a personal site on Neocities, very nice, but I realised that I needed a CMS. ProcessWire has been fabulous so far, while I've been creating files on my live hosting (LOL). I'm using Ai to help guide me through the "proper dev" practices, and honestly, it's been great. But now we're stuck. I'm trying to load my site locally. I've been getting various errors, blank screens. Ai insists that I need bootstrap.php and I need to get a fresh install of PW. When I look at the files in PW, there is no bootstrap.php. This page https://processwire.com/docs/front-end/include/ mentions bootstrapping, but no mention of the "missing" bootstrap.php file. Any ideas?

Nice! Thanks for the quick implementation! I'll give it a try and test a bit.

hi, wow, being french, a night later, i was about to say well the solution would be to create a custom field relying on pw getSelectablePages and something like tom select that allows lazy loading in the select but what @Jonathan Lahijani made is simply awesome!!! have a nice day

I must say, I really like this field. Wow... I'm impressed with GPT 5.5 and Codex CLI.

I made a few LOT more updates and I think it's ready to share for testing. I'm attaching it here as a zip for now (v88), but will publish to GitHub when I think it's ready. Edit: I have not really tested this as far as the configuration options go. My focus was mainly on functionality and UX. Obviously, don't use this on a production site, only test sites for now. Edit: it now works well in both Konkat (aka Default) and Core (aka Original) admin theme. InputfieldPageModalSelect-88.zip

Another verge large update with all the UX stuff: https://youtu.be/Xu6O84oUQKU

@jploch Good idea! It's been added.

@DrQuincy Here's a video of a concept field I made called InputfieldPageModalSelect. Let me know your thoughts and suggestions!

I agree with this. A new type of multi-select inputfield for page references for advanced use cases would definitely help. I know Shopify has some unique, purpose-built multi-select fields (like when choosing products/variations). But it doesn't let you search options that have been specifically selected. I didn't see anything generic from CMSes like Payload, Statamic, CraftCMS, Kirby or Laravel Filament. Have you seen anything in any other systems that solves this well?