I’ve used Obsidian for a number of years for my note-taking. I like the graph view, bidirectional linking, and its plain-text philosophy. For a while, nonetheless, one thing has been nagging me: its major supply of content material is typing phrases, and numerous them.
Just lately, I mounted this by pairing my Obsidian account with an area AI picture generator. I am not speaking about dropping in inventory pictures or copy-pasting content material from Midjourney or Leonardo.ai. I imply producing photos straight tied to my notes that dwell domestically, privately, and sure, with out a subscription.
The result’s a PKM (private data administration) workflow that lastly feels prefer it lives in either side of my mind.
OS
Home windows, macOS, Linux, Android, iOS, iPadOS
Developer
Dynalist Inc.
Why visuals matter in your Obsidian note-taking system
The science behind combining photos and textual content in your notes
I’m not a purely verbal thinker, regardless of writing for many of my skilled life. Once I’m mapping out a posh mission, constructing a personality for a narrative, or attempting to internalize an summary idea, a well-placed picture does one thing that one other paragraph or bullet factors can not. This picture can anchor the thought and provides my mind a deal with.
Including visuals the normal method, by internet searches, inventory websites, and screenshots, by no means felt proper to me. It felt much less private than I wished. I discovered them to be another person’s interpretation of the idea, not mine. And managing these picture recordsdata over time was all the time a time-waster.
What I wanted was a approach to generate the picture I had in my head, immediately, with out leaving my workflow.
arrange a free native AI picture generator in your Mac
Selecting the best Secure Diffusion mannequin in your {hardware}
Credit score: Bryan Wolfe / MakeUseOf
As a primarily Mac consumer, the instrument that made this click on was Diffusers, an open-source image-generation app from Hugging Face. It is free, obtainable for obtain within the Mac App Retailer, and extremely simple to put in with no command-line data in anyway. It runs totally on my machine and leverages Apple Silicon’s Metallic GPU, which implies it generates quicker than you’d anticipate from shopper {hardware}.
Once you first open it, you will be prompted to obtain a mannequin, which is the inventive engine that powers the picture technology. Should you’re on an M3 or M4 Mac with 16GB or extra of reminiscence, I would counsel going straight to Secure Diffusion 3 Medium. It produces one of the best photorealistic outcomes of something obtainable within the app, and your {hardware} will deal with it comfortably. Should you personal an older Mac, Secure Diffusion 2.1 is a dependable place to begin.
Should you’re on Home windows, there are a number of alternate options value contemplating. This contains ComfyUI, which presents a node-based visible interface and a one-click installer, and Automatic1111, which presents a extra conventional internet UI. Each are wonderful alternate options, however neither is optimized for Apple Silicon the best way Diffusers is. Mac customers will get noticeably higher efficiency staying within the Hugging Face ecosystem.
The important thing requirement for all of those options is that they run domestically. As such, your notes and concepts stay non-public. Though you may route them by a cloud service, that defeats the entire objective of a private data vault.
join Obsidian and Diffusers: a step-by-step workflow
Utilizing Obsidian’s Templater plugin to construct your picture prompts
Credit score: Bryan M. Wolfe / MakeUseOf
As soon as Diffusers is operating, I wanted a approach to combine it easily with Obsidian slightly than treating them as two separate apps that I went backwards and forwards between. Here is the workflow I landed on.
First, proceed to put in writing your concepts and ideas on Obsidian as you have all the time executed. I all the time make certain to put in writing at the least a paragraph in regards to the idea or node I am engaged on. The objective right here is for the writing to power me to articulate what the picture ought to seize, which makes the eventual immediate dramatically higher. In case your immediate is half-baked, the picture can be too.
Subsequent, use a template to construct your visible immediate. For this, I exploit Obsidian’s Templater plugin, one of many best ones obtainable. If you have not put in it, go to Settings -> Neighborhood Plugins -> Browse and seek for Templater, and set up it. It is free and takes about two minutes to arrange.
As soon as it is written, create a brand new folder in your vault referred to as Templates, and inside it, create a brand new word referred to as Picture Immediate Template.
I am at present writing a homicide thriller set on a cruise ship, so I’ve spent numerous time visualizing key characters. This implies documenting the “who” and the “what,” in addition to the temper, environment, and magnificence. I additionally embody a ‘destructive immediate’ to inform Diffusers what to actively keep away from (like blurry limbs or cartoonish types).
Associated
Is Obsidian Actually Definitely worth the Studying Curve for Notice-Taking?
Is Obsidian the final word note-taking instrument or simply an overcomplicated app?
For my first immediate, I’ve included this language: “Topic: Temper/environment: Colours: Type: hyper-detailed, cinematic. Adverse immediate:”
Now, every time I am able to create a picture for a word, I can open that word, click on inside it, and set off it with a hotkey. You may set this set off in Templater’s settings underneath Template Hotkeys. A small window pops up, and I kind “Picture” to seek out my template, then hit Enter. The template provides that construction to my word, and I merely fill within the blanks.
Here is an instance of a immediate that I’ve used not too long ago:
Topic: a retired schoolteacher in her mid-60s, silver hair pulled again loosely, studying glasses pushed up on her brow, sitting at a kitchen desk with a cup of tea
Temper/environment: quietly content material, unhurried, the sensation of a Sunday morning with nowhere to be
Colours: heat cream, delicate morning gentle, muted sage inexperienced
Type: hyper-detailed, cinematic, photorealistic
For a last step, you need to copy your accomplished immediate from Obsidian and convey it to Diffusers by pasting it into the Immediate textual content field on the prime of the app. Add your destructive immediate within the Adverse Immediate field beneath it. On this instance, I used:
Adverse immediate: blurry, low high quality, cartoon, watermark, textual content, deformed arms, younger, glamorous, dramatic lighting
As soon as each prompts are crammed in, select Generate. I normally do that two or thrice to seek out the composition I need. Imagine me, a few of the photos that Diffusers generates aren’t sound, so you will have to view a number of variations earlier than you discover the one you need. Over time, as you get higher at writing prompts, the method goes faster.
You now want to avoid wasting your picture to an property folder inside your Obsidian vault. For simplicity, I used “/property/photos/” in my Obsidian vault root. Each time you save a generated picture, it goes right here. On this instance, I named the file “schoolteacher.png.”
The ultimate step is to embed the picture on the prime of your Obsidian word. On this case, I would add “![[schoolteacher.png]]”. On first open, it orients me instantly, and now I do know what this word seems like earlier than I learn a single phrase.
The inventive payoff: what modifications when your notes have visuals
The connection I’ve established between Obsidian and Diffusers has yielded a number of advantages. First, it has modified how deeply I have interaction with particular person notes. There is a sense of authorship {that a} clean Markdown file with a bullet listing would not carry. It has additionally modified how I navigate my vault. The Obsidian Kanban and Dataview plugins can floor embedded photos in card and desk views, which turns a visible scan of my vault into one thing nearer to shopping {a magazine} than studying a spreadsheet. My graph view exhibits how concepts join; my image-forward views present what these concepts are. For inventive tasks particularly, this will really be transformative.
I will be the primary to confess: this setup did not essentially make me extra productive in a standard sense. It did not assist me write quicker or seize extra notes. Slightly, it made these notes higher over the long term. I additionally would not take into account this a “deep analysis mission,” which Obsidian is thought to assist many with.
Should you’ve ever felt like your notes are too flat, too text-heavy, too disconnected from the way you really suppose, that is well worth the afternoon. Your vault would not must be a submitting cupboard. With an area picture generator and some minutes of prompting, it will possibly really feel far more like a thoughts.
