> May 23, 2024 — We will be switching Fabric to Go in a few weeks to avoid all the installation issues with Python. The Go version will be dead-simple to install and will be even faster. Plus easier to update. We already have it working thanks to the heroic efforts of @xssdoctor, and we're just working on testing now! Stay tuned for more info on the release date!
This is a **brilliant** video by Network Chuck that goes over why he's started using Fabric for all things AI. He talks about the spirit of the project, how to install it, and how he uses it, and he just generally articulates the spirit of what we're doing here SO WELL. Thanks to Chuck for this!
Since the start of 2023 and GenAI we've seen a massive number of AI applications for accomplishing tasks. It's powerful, but _it's not easy to integrate this functionality into our lives._
<divalign="center">
<h4>In other words, AI doesn't have a capabilities problem—it has an <em>integration</em> problem.</h4>
</div>
Fabric was created to address this by enabling everyone to granularly apply AI to everyday challenges.
## Philosophy
> AI isn't a thing; it's a _magnifier_ of a thing. And that thing is **human creativity**.
We believe the purpose of technology is to help humans flourish, so when we talk about AI we start with the **human** problems we want to solve.
### Breaking problems into components
Our approach is to break problems into individual pieces (see below) and then apply AI to them one at a time. See below for some examples.
Prompts are good for this, but the biggest challenge I faced in 2023——which still exists today—is **the sheer number of AI prompts out there**. We all have prompts that are useful, but it's hard to discover new ones, know if they are good or not, _and manage different versions of the ones we like_.
One of <code>fabric</code>'s primary features is helping people collect and integrate prompts, which we call _Patterns_, into various parts of their lives.
Fabric has Patterns for all sorts of life and work activities, including:
Fabric _Patterns_ are different than most prompts you'll see.
- **First, we use `Markdown` to help ensure maximum readability and editability**. This not only helps the creator make a good one, but also anyone who wants to deeply understand what it does. _Importantly, this also includes the AI you're sending it to!_
- **Next, we are extremely clear in our instructions**, and we use the Markdown structure to emphasize what we want the AI to do, and in what order.
- **And finally, we tend to use the System section of the prompt almost exclusively**. In over a year of being heads-down with this stuff, we've just seen more efficacy from doing that. If that changes, or we're shown data that says otherwise, we will adjust.
The most feature-rich way to use Fabric is to use the `fabric` client, which can be found under <ahref="https://github.com/danielmiessler/fabric/tree/main/installer/client">`/client`</a> directory in this repository.
Ensure you have at least python3.10 installed on your operating system. Otherwise, when you attempt to run the pip install commands, the project will fail to build due to certain dependencies.
If you want to use it with OpenAI API-compatible inference servers, such as [FastChat](https://github.com/lm-sys/FastChat), [Helmholtz Blablador](http://helmholtz-blablador.fz-juelich.de), [LM Studio](https://lmstudio.ai) and others, simply export the following environment variables:
The client, by default, runs Fabric patterns without needing a server (the Patterns were downloaded during setup). This means the client connects directly to OpenAI using the input given and the Fabric pattern used.
1. Run the `summarize` Pattern based on input from `stdin`. In this case, the body of an article.
```bash
pbpaste | fabric --pattern summarize
```
2. Run the `analyze_claims` Pattern with the `--stream` option to get immediate and streaming results.
3. Run the `extract_wisdom` Pattern with the `--stream` option to get immediate and streaming results from any Youtube video (much like in the original introduction video).
If you're not looking to do anything fancy, and you just want a lot of great prompts, you can navigate to the [`/patterns`](https://github.com/danielmiessler/fabric/tree/main/patterns) directory and start exploring!
We hope that if you used nothing else from Fabric, the Patterns by themselves will make the project useful.
You can use any of the Patterns you see there in any AI application that you have, whether that's ChatGPT or some other app or website. Our plan and prediction is that people will soon be sharing many more than those we've published, and they will be way better than ours.
But we go beyond just providing Patterns. We provide code for you to build your very own Fabric server and personal AI infrastructure!
## Structure
Fabric is themed off of, well… _fabric_—as in…woven materials. So, think blankets, quilts, patterns, etc. Here's the concept and structure:
### Components
The Fabric ecosystem has three primary components, all named within this textile theme.
- The **Mill** is the (optional) server that makes **Patterns** available.
- **Patterns** are the actual granular AI use cases (prompts).
- **Stitches** are chained together _Patterns_ that create advanced functionality (see below).
- **Looms** are the client-side apps that call a specific **Pattern** hosted by a **Mill**.
### CLI-native
One of the coolest parts of the project is that it's **command-line native**!
Each Pattern you see in the `/patterns` directory can be used in any AI application you use, but you can also set up your own server using the `/server` code and then call APIs directly!
One key feature of `fabric` and its Markdown-based format is the ability to _directly reference_ (and edit) individual [Patterns](#components) directly—on their own—without any surrounding code.
Even better, you can also have your [Mill](#components) functionality directly call _system_ and _user_ prompts from `fabric`, meaning you can have your personal AI ecosystem automatically kept up to date with the latest version of your favorite [Patterns](#components).
Here's an abridged output example from the <ahref="https://github.com/danielmiessler/fabric/blob/main/patterns/extract_wisdom/system.md">`extract_wisdom`</a> pattern (limited to only 10 items per section).
```bash
# Paste in the transcript of a YouTube video of Riva Tez on David Perrel's podcast
pbpaste | extract_wisdom
```
```markdown
## SUMMARY:
The content features a conversation between two individuals discussing various topics, including the decline of Western culture, the importance of beauty and subtlety in life, the impact of technology and AI, the resonance of Rilke's poetry, the value of deep reading and revisiting texts, the captivating nature of Ayn Rand's writing, the role of philosophy in understanding the world, and the influence of drugs on society. They also touch upon creativity, attention spans, and the importance of introspection.
## IDEAS:
1. Western culture is perceived to be declining due to a loss of values and an embrace of mediocrity.
2. Mass media and technology have contributed to shorter attention spans and a need for constant stimulation.
3. Rilke's poetry resonates due to its focus on beauty and ecstasy in everyday objects.
4. Subtlety is often overlooked in modern society due to sensory overload.
5. The role of technology in shaping music and performance art is significant.
6. Reading habits have shifted from deep, repetitive reading to consuming large quantities of new material.
7. Revisiting influential books as one ages can lead to new insights based on accumulated wisdom and experiences.
8. Fiction can vividly illustrate philosophical concepts through characters and narratives.
9. Many influential thinkers have backgrounds in philosophy, highlighting its importance in shaping reasoning skills.
10. Philosophy is seen as a bridge between theology and science, asking questions that both fields seek to answer.
## QUOTES:
1. "You can't necessarily think yourself into the answers. You have to create space for the answers to come to you."
NEW FEATURE! We have incorporated [PraisonAI](https://github.com/MervinPraison/PraisonAI) into Fabric. This feature creates AI agents and then uses them to perform a task.
These are helper tools to work with Fabric. Examples include things like getting transcripts from media files, getting metadata about media, etc.
## yt (YouTube)
`yt` is a command that uses the YouTube API to pull transcripts, pull user comments, get video duration, and other functions. It's primary function is to get a transcript from a video that can then be stitched (piped) into other Fabric Patterns.
```bash
usage: yt [-h] [--duration] [--transcript] [url]
vm (video meta) extracts metadata about a video, such as the transcript and the video's duration. By Daniel Miessler.
'ts' is a command that uses the OpenAI Whisper API to transcribe audio files. Due to the context window, this tool uses pydub to split the files into 10 minute segments. for more information on pydub, please refer https://github.com/jiaaro/pydub
`save` is a "tee-like" utility to pipeline saving of content, while keeping the output stream intact. Can optionally generate "frontmatter" for PKM utilities like Obsidian via the
"FABRIC_FRONTMATTER" environment variable
If you'd like to default variables, set them in `~/.config/fabric/.env`. `FABRIC_OUTPUT_PATH` needs to be set so `save` where to write. `FABRIC_FRONTMATTER_TAGS` is optional, but useful for tracking how tags have entered your PKM, if that's important to you.
save: a "tee-like" utility to pipeline saving of content, while keeping the output stream intact. Can optionally generate "frontmatter" for PKM utilities like Obsidian via the
"FABRIC_FRONTMATTER" environment variable
positional arguments:
stub stub to describe your content. Use quotes if you have spaces. Resulting format is YYYY-MM-DD-stub.md by default
options:
-h, --help show this help message and exit
-t, TAG, --tag TAG add an additional frontmatter tag. Use this argument multiple timesfor multiple tags
-n, --nofabric don't use the fabric tags, only use tags from --tag
-s, --silent don't use STDOUT for output, only save to the file
- _Jason Haddix_ for the idea of a stitch (chained Pattern) to filter content using a local model before sending on to a cloud model, i.e., cleaning customer data using `llama2` before sending on to `gpt-4` for analysis.
- _Dani Goland_ for enhancing the Fabric Server (Mill) infrastructure by migrating to FastAPI, breaking the server into discrete pieces, and Dockerizing the entire thing.
