Видео

Nice video, but to me this plugin is not necessary. In a daily not template, I simply have task queries, to pull up tasks due that day, overdue, due in 2 days, etc. And each task by default, via the tasks plugin, comes with an edit button. As well as a quick 'defer to tomorrow button. So no matter where the task originated from, I will always see every task on every day I am supposed to, and, can also edit every task. It is very easy to setup, the task queries are simple and allows much for 'at a glance' various views of your tasks than this roll over. Everyone has their way of doing things. I am not shooting this rollover idea down if it works for you and others. I just wanted to point out there are other ways of doing it, which, in my opinion are better. Cheers!

TG Error text is not a function

If anyone is wondering about the theme, it's Minimal theme, get it and get it's extension (minimal) and then from it's options choose Rose Pine

It seems like you were expecting it to work similarly to Obsidian going into it? But why? Obsidian is more power-user software, whereas it seems Reflect is geared towards a mainstream userbase who is used to the world of apps rather than markdown files and plug-ins.

bro use Command + ~ or Command + shift + tab

stop using AI. The c02, water and energy is horrifying

What a interesting idea. Any recommendations?

This isn't what my System Sculpt looks like? is this unreleased?

hey, where i can find this Janitor plugin? i would need that!

I would prefer a customizable limit for context files. Gemini has one to two million tokens context limit, so it should be able to handle thousands of files. You’re doing great work, keep it up!

I'm trying to learn vim and I've been using Obsidian a lot lately so I enabled vim in Obsidian as a way to practice the muscle memory I guess. But idk it's just not working for my Obsidian workflow. The way I use Obsidian involves a lot of moving around to different notes, and when I'm in a specific note I don't move around within it all that much, since my notes are pretty short and don't have a ton of sections. But we'll see, maybe I'll get a lot of use with the replacing characters and such

these are some good ideas

Great content! I think it would be an awesome feature to allow the user to select obsidian notes as prompts You already operating in (IMHO) the best markdown editor that exists, you might as well use that power and functionality. This could also open the door to much more powerful dynamic prompt templates while you as the SystemSculpt dev wouldn't have to add it in manually. I'm imagining using the power of obsidian note templates and templater to generate prompts using predefined prompt template templates and dynamically filling in the real prompt template based on whatever templater config you have set up This way you could generate generic system prompt templates for various tasks but then allow for much more specific template instructions and formatting for different use cases which could be configured via templater (folder template, document/note type template, etc)

I think it would be beneficial to allow users to save custom templates so they can only have to go through the work once of making sure the variables are right once This could actually be done with a slightly different workflow where you duplicate a model provider config and then start editing the duplicate. This way, everything is the same by default, thereby reducing user-error. The user would only have to change the relevant fields (model name, other model-specific params) as needed Thanks for all the great work! Loving the new posts for 2025, please keep it up!

As you noted in the video, the notification for the 100 file limit was hard to catch. This might lead to someone thinking the model now has the context of all 1000 files instead of the limit of 100 Maybe change the alert for that to stay up for much longer than the others so the user sees it

Subscribed! Looking forward to this series.

Maybe a “power user” mode and a “casual user” mode? Power users can have more freedom to custom tabs casual users have more out of the box one click options. IMO…

Can we use local custom Whisper models for transcriptions?

Ref yor question, Currently for custom Providers I have OpenAI, Groq, Open Router and my Local endpoint is pointed at Ollama. I have no plans to use LM Studio. However because its list of models is so comprehensive and prices so competative I would be quite happy with just OpenRouter. On the later screen about what Local models were available what I missed was some sort of guidance on what a particular model was better for.

it looks great to me

I feel there should be 2 to 6 models in each of your three categories because aa choice of two seems a minimum and a max of six is what fits on the screen and more would just be confusing. You did not explain how the plug in is getting to these models you are curating and how use is paid for. In the custom set up one either uses an APi to (say) OpenRouter or goes direct to local e.g with Ollama. Ref Ollama it would be great to have help in a similar way for choosing which LLMs to use as I dont find that easy at all and a system like you described today would be great for that. Issue then can be that you might recomend a model but t will not be available unless the appropriate ollama command has been issued. It would be great if there was a way around that bottlenecjk so if one choose an ollama model that isnt yet downloaded it gets downloaded. Ollama File Procesing Currently I am experimenmting with using Ollama calls from Python to process project files. All my notes are in one folder so to select the files to process I use Python to select only files with Project:: <projectName> at the top. Then I go through all the files applying a prompt to alll the files to extract the required info. At the moment I cant do this in SystemSculpt AI because there is no facility to cycle through files based on containing certain text - in this case "Project:: .... ". Maybe you could consider adding that at some point. The aim of my current project is to process all my History Files to produce a Timeline with links to the source files. This wold be expensive with an online Model but with a combination of Python and Ollama it can be set going and it does not matter if it takes several minutes to complete.

As much as I do find these videos kinda weird I do like tuning in every day

Love the content. There is an issue with your website. I would like to point out that currently if you have Avast installed, you can't navigate to your site. Web Shield displays Connection aborted. Threat URL:Mal. This will cause people to stop using the plugin when they click the documentation and are redirected to your site.

Like ShaneHolloman, I'm also not seeing any menu items when I right click. And my UI is different from that in the video. It shows up in the sidebar. The task, recorder and chat buttons are displayed in the status bar, not next to the input box. What version is this one in the video? Mine is 0.7.0 installed in Windows.

Subscribed! This looks promising. Can't wait to play with this. Thanks for developing this plugin!

Is there an add-in available for TextAloud markup files?

I feel like these videos are hypnotizing/brainwashing

Can this work with a local instance of whisper?

heya... are these upcoming features? I only ask as I downloaded the latest available (0.7.0) and right clicking mp3 doesn't have the options you do. Just has a `SystemSculpt - Transcribe` option. I use olllama only. Maybe that's why I don't have the same options available that you do? Solid work btw ...

interesting but doesn't sync and Idon't know why, what a pity

Nah bro knew what he was doing with these images

Very impressive. Ref limits - isnt it hard to have a general way to have limits after all there are limits imposed by the model one is using eithere by its token limit or the cost of tokens and different people will have a different approach to that. For myself I remember trying to do a one off task once where I wanted to analyse a set of insurance pdf documents to see if they complied to some guildlines for what sort of insurance our type of charity should have. On this occasion I did not care about the cost as it was a one off. (Cost was only about 30c anyway). So I would say needs to be flexible but easy to use. Another thing that I am thinking of trying to do in Obsidian somehow aided by a local LLM with Ollama is to produce a master notes index. I have a sort of system working using front matter and tags and Dataview but it would be lovely if somehow there was a way to have a sort of auto indexing that could just do it in the background based on the contents of notes.

I'd really like to see local whisper integration like having Buzz built in so I can keep voice transcription local.

Fantastic. It would be great to be able to establish a light-graphRAG in the system to be able to augment the identification of links and themes in the system.

My friend I love the work, but you need to add support for custom AI providers. I use a lot of different providers to test and utilise them. For example, Gemini 1.5 and 2.0 power all my clipping tasks due to their massive context windows, which are very useful for RUclips. Groq is great for quick summaries because of its speed, and Deepseek is ideal for coding. Currently, I can only see three providers available in your plugin, and this is a dealbreaker for me to start using it.

Might be nice to have a token limit also, for just the use case of small files that you mentioned.

Nice, keep us posted . Love to try this out. How does this compare to the Obsidian plugin copilot were you index your whole vault with an embedded model (vector based) after which you can ask specific question on all of your notes?

👍

This is some neurotic control shit that rich people tell themself is for "efficiency" 😂

looking forward to this!

This looks fantastic. Please bring us all with you. I really need to integrate AI into my vault but I've struggled to find my way around your videos so i got nowhere in 2024. I feel like there's a great system here, just waiting to burst out, like an infant Xenomorph.

My favorite AI everything for obsidian. Excellent work!

Fabulous overview! New to LLM implementation, and "Creative" vs "Analytic" is exactly what I needed to know, even though that's not what I searched to get your video lol! I need both for my project, but for different applications. You gave me exactly what I needed. Thank you!

thanks a million!

对于喜欢美化但不会写代码的人来说，你做的这个教程简单易懂，非常感谢！

In the MacOS by default to go forward, you can use the command + tab and then the tilde symbol. I like your video and have already installed the application mentioned. Just because on a Windows PC, you press ALT + TAB to see all opened windows, so if you have 2 browsers, you see both with preview, and you know what to choose. But on MacOS, by default, you see only one browser icon and switch between the last active window.

whats the theme you are using?

You should seperate format prompts. Also for summeries I recommend to reduce the token number quite a bit. Test what works for your style and also specify the bullet point number.

The name is just git of this plugin. But seems to be the same.

how to automatically push images to github using this extension?