├── .markdownlint.json
├── src
├── content
│ └── en
│ │ ├── search.md
│ │ ├── featured-background.png
│ │ ├── docs
│ │ ├── Workflows
│ │ │ ├── Creating Photorealistic Images
│ │ │ │ ├── duckies.png
│ │ │ │ └── _index.md
│ │ │ ├── _index.md
│ │ │ ├── Creating Presentations from scratch
│ │ │ │ └── _index.md
│ │ │ └── Creating a annotated speaker deck for training
│ │ │ │ └── _index.md
│ │ ├── _index.md
│ │ ├── AI Tools
│ │ │ ├── chatgpt
│ │ │ │ └── _index.md
│ │ │ ├── _index.md
│ │ │ ├── whisper
│ │ │ │ └── _index.md
│ │ │ ├── midjourney
│ │ │ │ └── _index.md
│ │ │ └── slackgpt
│ │ │ │ └── _index.md
│ │ └── Responsible AI
│ │ │ └── _index.md
│ │ └── _index.md
├── layouts
│ └── 404.html
├── go.mod
├── .gitignore
├── assets
│ └── scss
│ │ └── _variables_project.scss
├── CONTRIBUTING.md
├── package.json
├── go.sum
├── hugo.toml
└── LICENSE
├── .vscode
└── settings.json
├── .github
├── dependabot.yml
└── workflows
│ └── hugo.yml
├── README.md
├── .devcontainer
├── devcontainer.json
└── Dockerfile
└── .gitignore
/.markdownlint.json:
--------------------------------------------------------------------------------
1 | {
2 | "MD013": false,
3 | "MD022": false
4 | }
5 |
--------------------------------------------------------------------------------
/src/content/en/search.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Search Results
3 | layout: search
4 | ---
5 |
--------------------------------------------------------------------------------
/src/content/en/featured-background.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/xebia/ai-toolbox/HEAD/src/content/en/featured-background.png
--------------------------------------------------------------------------------
/src/content/en/docs/Workflows/Creating Photorealistic Images/duckies.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/xebia/ai-toolbox/HEAD/src/content/en/docs/Workflows/Creating Photorealistic Images/duckies.png
--------------------------------------------------------------------------------
/src/layouts/404.html:
--------------------------------------------------------------------------------
1 | {{ define "main" -}}
2 |
3 |
Not found
4 |
Oops! This page doesn't exist. Try going back to the home page.
5 |
6 | {{- end }}
7 |
--------------------------------------------------------------------------------
/.vscode/settings.json:
--------------------------------------------------------------------------------
1 | {
2 | "cSpell.words": [
3 | "Midjourney",
4 | "Xebia"
5 | ],
6 | "grammarly.selectors": [
7 | {
8 | "language": "markdown",
9 | "scheme": "file"
10 | }
11 | ]
12 | }
13 |
--------------------------------------------------------------------------------
/src/go.mod:
--------------------------------------------------------------------------------
1 | module github.com/google/docsy-example
2 |
3 | go 1.12
4 |
5 | require (
6 | github.com/FortAwesome/Font-Awesome v0.0.0-20230327165841-0698449d50f2 // indirect
7 | github.com/google/docsy v0.7.0 // indirect
8 | github.com/twbs/bootstrap v5.2.3+incompatible // indirect
9 | )
10 |
--------------------------------------------------------------------------------
/src/.gitignore:
--------------------------------------------------------------------------------
1 | ### Hugo ###
2 | # Generated files by hugo
3 | /public/
4 | /resources/_gen/
5 | /assets/jsconfig.json
6 | /hugo_stats.json
7 |
8 | # Executable may be added to repository
9 | hugo.exe
10 | hugo.darwin
11 | hugo.linux
12 |
13 | # Temporary lock file while building
14 | /.hugo_build.lock
15 |
16 | # Go workspace file
17 | go.work
--------------------------------------------------------------------------------
/.github/dependabot.yml:
--------------------------------------------------------------------------------
1 | # To get started with Dependabot version updates, you'll need to specify which
2 | # package ecosystems to update and where the package manifests are located.
3 | # Please see the documentation for all configuration options:
4 | # https://docs.github.com/github/administering-a-repository/configuration-options-for-dependency-updates
5 |
6 | version: 2
7 | updates:
8 | - package-ecosystem: "github-actions" # See documentation for possible values
9 | directory: "/" # Location of package manifests
10 | schedule:
11 | interval: "weekly"
12 |
--------------------------------------------------------------------------------
/src/assets/scss/_variables_project.scss:
--------------------------------------------------------------------------------
1 | /*
2 |
3 | Add styles or override variables from the theme here.
4 |
5 | */
6 |
7 | $primary: #6a1d57;
8 |
9 | $btn-bg: #ffffff;
10 | $btn-secondary: #ffffff;
11 | $secondary: #ffffff;
12 |
13 | :root {
14 | .text-info {
15 | --bs-info-rgb: 255,255,255
16 | }
17 |
18 | .td-overlay--dark::after{
19 | background: rgba(0,0,0,0.3);
20 | }
21 |
22 | .mt-md-5{
23 | margin-top: 15rem !important;
24 | }
25 |
26 | .td-box--2 {
27 | background: rgba(247,248,252,1)
28 | }
29 | }
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # Xebia AI Toolbox
2 | > Check out the website [here][1]
3 |
4 | There are a lot of AI tools, but how do you actually accomplish any work with them? This project aims to collect as many of your workflows as you want to share with fellow Xebians to help them to be more productive using AI.
5 | Found a way to capture your thoughts with AI? Write about it! Does it help you with blogging/speaking/writing? Write about it!
6 |
7 | ## How to contribute
8 | Start a new branch, open a Codespace and start in `src/docs/`. You can add a new `_index.md` file and start writing. Alternatively, you can clone this repo locally, but a Codespace comes with all the tools installed.
9 |
10 | If you want to test your changes, you can run:
11 |
12 | ```sh
13 | cd src/
14 | hugo server
15 | ```
16 |
17 | If you like your changes, simply shoot in a PR!
18 |
19 | [1]: https://xebia.github.io/ai-toolbox/
20 |
--------------------------------------------------------------------------------
/.devcontainer/devcontainer.json:
--------------------------------------------------------------------------------
1 | {
2 | "name": "Hugo (Community)",
3 | "build": {
4 | "dockerfile": "Dockerfile",
5 | "args": {
6 | "VARIANT": "hugo_extended",
7 | "VERSION": "latest",
8 | "NODE_VERSION": "14"
9 | }
10 | },
11 |
12 | "customizations": {
13 | "vscode": {
14 | "settings": {
15 | "html.format.templating": true
16 | },
17 |
18 | "extensions": [
19 | "bungcip.better-toml",
20 | "davidanson.vscode-markdownlint",
21 | "rusnasonov.vscode-hugo",
22 | "github.copilot",
23 | "streetsidesoftware.code-spell-checker",
24 | "fivethree.vscode-hugo-snippets",
25 | "eamodio.gitlens",
26 | "yzhang.markdown-all-in-one",
27 | "znck.grammarly"
28 | ]
29 | }
30 | },
31 |
32 | "forwardPorts": [
33 | 1313
34 | ],
35 |
36 | "postCreateCommand": "git config --global --add safe.directory /workspace/ai-toolbox",
37 |
38 | "remoteUser": "node"
39 | }
40 |
--------------------------------------------------------------------------------
/src/content/en/docs/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Xebia AI Toolbox Documentation
3 | linkTitle: Home
4 | menu: {main: {weight: 20}}
5 | weight: 20
6 | ---
7 |
8 | We believe AI can help us if we use it in a responsible way. This means we are careful when working with AI tools regarding client data, privacy and security. We also believe in the power of open source and sharing knowledge. That is why we created this site to share our knowledge and help eachother in our daily work.
9 |
10 | This website consits of a couple of parts: Workflows & Tools. The workflows are the actual workflows you can use to get things done. The tools are the tools you can use to get things done. The tools are often used in the workflows.
11 |
12 | Next to using tools and workflows we also want to share our knowledge on how to use AI in a responsible way. This is why we have a section on Responsible AI where we explain how to use AI in such a way that it doesn't hurt our company or our customers.
13 |
--------------------------------------------------------------------------------
/src/CONTRIBUTING.md:
--------------------------------------------------------------------------------
1 | # How to Contribute
2 |
3 | We'd love to accept your patches and contributions to this project. There are
4 | just a few small guidelines you need to follow.
5 |
6 | ## Code reviews
7 |
8 | All submissions, including submissions by project members, require review. We
9 | use GitHub pull requests for this purpose. Consult
10 | [GitHub Help](https://help.github.com/articles/about-pull-requests/) for more
11 | information on using pull requests.
12 |
13 | ## Content guidelines
14 |
15 | Make sure all content you contribute is in a readable state. Pages should have enough content to be useful but you don't have to write all the details if you don't have time or knowledge to add that. We can always add more content later.
16 |
17 | If content / guides on using certain tools is found by other creators on the internet please do add a link back to them. We don't want to steal content from others, this website is just to empower everyone within Xebia to use these tools and provide a central place to find information on it.
--------------------------------------------------------------------------------
/src/content/en/docs/Workflows/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: AI Workflows
3 | weight: 1
4 | categories: [Workflows]
5 | tags: [Workflows]
6 | description: >
7 | AI Workflow Collection
8 | ---
9 |
10 | This website focuses on workflows to actually get things done instead of tools. This section describes a number of workflows to be more productive using AI. Have your own workflow of being more efficient & productive? please add it to this website!
11 |
12 | ## Responsible AI Checklist
13 | As we think responsible AI is essential, we should always consider this when using AI tools.
14 | Add this checklist to every workflow/tool page you create so people have a quick overview of what to consider when using this workflow or tool.
15 |
16 | |Responsible AI Checklist|
17 | |---|---|
18 | |Can be used with customer data| ℹ️ Ask Consent / ✅ Yes, this tool is safe to use / ❌ No, this tool is not safe to use|
19 | |Can the output be used at customer| ✅ Yes / ❌ No |
20 | |Can be used with Xebia Internal secret data| ✅ Yes, this tool is safe to use / ❌ No, this tool is not safe to use|
21 | |Can the output be used commercially by Xebia| ✅ Yes / ❌ No |
22 | |Is data being stored in the region you use it | ✅ Yes / ❌ No|
23 | |Do we have a Xebia license / managed solution for this tool | ✅ Yes / ❌ No|
24 | |Tools used in workflow| Links to tool pages for these tools |
25 |
--------------------------------------------------------------------------------
/.devcontainer/Dockerfile:
--------------------------------------------------------------------------------
1 | # Update the NODE_VERSION arg in docker-compose.yml to pick a Node version: 18, 16, 14
2 | ARG NODE_VERSION=16
3 | FROM mcr.microsoft.com/devcontainers/javascript-node:${NODE_VERSION}
4 |
5 | # VARIANT can be either 'hugo' for the standard version or 'hugo_extended' for the extended version.
6 | ARG VARIANT=hugo
7 | # VERSION can be either 'latest' or a specific version number
8 | ARG VERSION=latest
9 |
10 | # Download Hugo
11 | RUN apt-get update && apt-get install -y ca-certificates openssl git curl && \
12 | rm -rf /var/lib/apt/lists/* && \
13 | case ${VERSION} in \
14 | latest) \
15 | export VERSION=$(curl -s https://api.github.com/repos/gohugoio/hugo/releases/latest | grep "tag_name" | awk '{print substr($2, 3, length($2)-4)}') ;;\
16 | esac && \
17 | echo ${VERSION} && \
18 | case $(uname -m) in \
19 | aarch64) \
20 | export ARCH=ARM64 ;; \
21 | *) \
22 | export ARCH=64bit ;; \
23 | esac && \
24 | echo ${ARCH} && \
25 | wget -O ${VERSION}.tar.gz https://github.com/gohugoio/hugo/releases/download/v${VERSION}/${VARIANT}_${VERSION}_Linux-${ARCH}.tar.gz && \
26 | tar xf ${VERSION}.tar.gz && \
27 | mv hugo /usr/bin/hugo
28 |
29 | # Download Go
30 | ENV GOLANG_VERSION 1.20.5
31 | ENV GOPATH /go
32 | ENV PATH $GOPATH/bin:/usr/local/go/bin:$PATH
33 | RUN curl -fsSL "https://dl.google.com/go/go$GOLANG_VERSION.linux-amd64.tar.gz" | tar -C /usr/local -xz
34 |
35 | EXPOSE 1313
--------------------------------------------------------------------------------
/src/content/en/docs/AI Tools/chatgpt/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: ChatGPT
3 | weight: 4
4 | categories: [Tools]
5 | tags: [SlackGPT, ChatGPT]
6 | description: >
7 | ChatGPT is a chatbot powered by OpenAI's GPT models.
8 | ---
9 |
10 | |Responsible AI Checklist|
11 | |---|---|
12 | |Can be used with customer data| ❌ No|
13 | |Can the output be used at customer| ✅ Yes|
14 | |Can be used with Xebia Internal secret data| ❌ No|
15 | |Can the output be used commercially by Xebia| ❓TODO VERIFY PROOF |
16 | |Is data being stored in the region you use it | ❌ No|
17 | |Do we have a Xebia license / managed solution for this tool | ✅ Yes, please use [SlackGPT][1]|
18 |
19 | ## What is ChatGPT
20 | ChatGPT (Chat Generative Pre-Trained Transformer) is an artificial intelligence chatbot developed by OpenAI.
21 |
22 | ## How to use ChatGPT
23 | You can visit [the website][2] to have a conversation with ChatGPT.
24 |
25 | ## License / Costs
26 | While ChatGPT is free, you can opt for a ChatGPT Plus license. It enables the following:
27 |
28 | - General access to ChatGPT, even during peak times
29 | - Faster response times
30 | - Priority access to new features and improvements
31 |
32 | ## Suitable to use with clients?
33 | In general, no. ChatGPT stores all your conversations, including all the data you sent to ChatGPT. You can turn off this feature but still you're sending sensitive data to OpenAI.
34 |
35 | If you need a safer chat-like experience backed by a GPT model, try Xebia's [SlackGPT][1]. This application uses a paid endpoint, which ensures your data is not used to train the GPT models.
36 |
37 | [1]: /docs/ai-tools/slackgpt
38 | [2]: https://chat.openai.com/
39 |
--------------------------------------------------------------------------------
/src/content/en/docs/AI Tools/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: AI Tools
3 | weight: 4
4 | categories: [Tools]
5 | tags: [Tools]
6 | description: >
7 | AI Tools Collection
8 | ---
9 |
10 | This website focusses on workflows to actually get things done instead of tools. When using AI tools they often require licenses and some might not be suitable to work with client data. Are you using a specific tool create a page on a couple of things:
11 |
12 | - What is the tool?
13 | - how to use it?
14 | - What is the license?
15 | - Is it suitable to be used at clients
16 | - Is there a Xebia managed solution / alternative you could use?
17 | - If the tool can be used at clients how to convince your client to use it. what is the proof that this tool is safe to use, will not steal data, etc. Colleagues can use this proof to have conversations with clients.
18 |
19 | ## Responsible AI Checklist
20 |
21 | Add this checklist to every workflow / tool page you are creating so people have a quick overview of what to consider when using this workflow or tool.
22 |
23 | |Responsible AI Checklist|
24 | |---|---|
25 | |Can be used with customer data| ℹ️ Ask Consent / ✅ Yes this tool is safe to use / ❌ No this tool is not safe to use|
26 | |Can the output be used at customer| ✅ Yes / ❌ No |
27 | |Can be used with Xebia Internal secret data| ✅ Yes this tool is safe to use / ❌ No this tool is not safe to use|
28 | |Can the output be used commercially by Xebia| ✅ Yes / ❌ No |
29 | |Is data being stored in the region you use it | ✅ Yes / ❌ No|
30 | |Do we have a Xebia license / managed solution for this tool | ✅ Yes / ❌ No|
31 | |Tools used in workflow| Links to tool pages for these tools |
32 |
--------------------------------------------------------------------------------
/src/package.json:
--------------------------------------------------------------------------------
1 | {
2 | "name": "docsy-example-site",
3 | "version": "0.7.0-dev.0-unreleased",
4 | "description": "Xebia AI Toolbox",
5 | "repository": "github:xebia/ai-toolbox",
6 | "homepage": "https://ai-toolbox.xebia.dev",
7 | "author": "Xebia",
8 | "license": "Apache-2.0",
9 | "bugs": "https://github.com/xebia/ai-toolbox/issues",
10 | "spelling": "cSpell:ignore HTMLTEST precheck postbuild -",
11 | "scripts": {
12 | "_build": "npm run _hugo-dev",
13 | "_check:links": "echo IMPLEMENTATION PENDING for check-links; echo",
14 | "_hugo": "hugo --cleanDestinationDir",
15 | "_hugo-dev": "npm run _hugo -- -e dev -DFE",
16 | "_serve": "npm run _hugo-dev -- --minify serve",
17 | "build:preview": "npm run _hugo-dev -- --minify --baseURL \"${DEPLOY_PRIME_URL:-/}\"",
18 | "build:production": "npm run _hugo -- --minify",
19 | "build": "npm run _build",
20 | "check:links:all": "HTMLTEST_ARGS= npm run _check:links",
21 | "check:links": "npm run _check:links",
22 | "clean": "rm -Rf public/* resources",
23 | "make:public": "git init -b main public",
24 | "precheck:links:all": "npm run build",
25 | "precheck:links": "npm run build",
26 | "postbuild:preview": "npm run _check:links",
27 | "postbuild:production": "npm run _check:links",
28 | "serve": "npm run _serve",
29 | "test": "npm run check:links",
30 | "update:pkg:dep": "npm install --save-dev autoprefixer@latest postcss-cli@latest",
31 | "update:pkg:hugo": "npm install --save-dev --save-exact hugo-extended@latest"
32 | },
33 | "devDependencies": {
34 | "autoprefixer": "^10.4.14",
35 | "hugo-extended": "0.113.0",
36 | "postcss-cli": "^10.1.0"
37 | }
38 | }
39 |
--------------------------------------------------------------------------------
/src/content/en/docs/AI Tools/whisper/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Whisper AI
3 | weight: 4
4 | categories: [Tools]
5 | tags: [WhisperAI]
6 | description: >
7 | Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification. You can generate transcripts from audio files.
8 | ---
9 |
10 | |Responsible AI Checklist|
11 | |---|---|
12 | |Can be used with customer data| ❌ No|
13 | |Can the output be used at customer| ✅ Yes|
14 | |Can be used with Xebia Internal secret data| ❌ No|
15 | |Can the output be used commercially by Xebia| ❓TODO VERIFY PROOF |
16 | |Is data being stored in the region you use it | ❌ No|
17 | |Do we have a Xebia license / managed solution for this tool | ❌ No|
18 |
19 | ## What is Whisper AI
20 | Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification. You can generate transcripts from audio files.
21 |
22 | ## How to use Whisper AI
23 | You can visit [the website][1] to follow instructions. In short. You download the tools and run whisper on your commandline. You can use multiple language models. The bigger the model, the more accurate the transcript. Although in many cases, small or medium is sufficient.
24 |
25 | ## License / Costs
26 | whisper is free
27 |
28 | ## Suitable to use with clients?
29 | It generates a transcript from an audio file. It does so locally. But please check customer. Also make sure the transcript does not end up somewhere.
30 |
31 | [1]: https://github.com/openai/whisper
32 |
--------------------------------------------------------------------------------
/src/content/en/docs/AI Tools/midjourney/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Midjourney
3 | weight: 4
4 | categories: [Tools]
5 | tags: [Midjourney]
6 | description: >
7 | Midjourney is a generative AI that produces images.
8 | ---
9 |
10 | |Responsible AI Checklist|
11 | |---|---|
12 | |Can be used with customer data| ℹ️ Ask Consent|
13 | |Can the output be used at customer| ✅ Yes|
14 | |Can be used with Xebia Internal secret data| ❌ Any input you give Midjourney is publicly visible by default|
15 | |Can the output be used commercially by Xebia| ✅ Yes|
16 | |Is data being stored in the region you use it | ❌ No|
17 | |Do we have a Xebia license / managed solution for this tool | ❌ Yes|
18 |
19 | ## What is Midjourney
20 | From the website:
21 | > Midjourney is an independent research lab exploring new mediums of thought and expanding the imaginative powers of the human species.
22 |
23 | However, when we refer to Midjourney on this website, we mean the Midjourney Discord Bot. Currently, the Midjourney generative AI is only available through Discord. Interacting with the software is done by adding this bot to your Discord channel or chatting with the bot directly.
24 |
25 | ## How to use Midjourney
26 | The Quick Start is available [here][2].
27 |
28 | The Pro and Mega plan offer [Stealth Mode][3]. This mode prevents your generated content from publicly appearing on Midjourney's website. This mode is not enabled by default.
29 |
30 | ## License / Costs
31 | Midjourney has several [subscription plans][1]. From the docs:
32 | > You must purchase the Pro or Mega plan if you are a company making more than $1,000,000 USD in gross revenue per year.
33 |
34 | ## Suitable to use with clients?
35 | [1]: https://docs.midjourney.com/docs/plans
36 | [2]: https://docs.midjourney.com/docs/quick-start
37 | [3]: https://docs.midjourney.com/docs/stealth
38 |
--------------------------------------------------------------------------------
/src/go.sum:
--------------------------------------------------------------------------------
1 | github.com/FortAwesome/Font-Awesome v0.0.0-20220831210243-d3a7818c253f h1:bvkUptSRPZBr3Kxuk+bnWCEmQ5MtEJX5fjezyV0bC3g=
2 | github.com/FortAwesome/Font-Awesome v0.0.0-20220831210243-d3a7818c253f/go.mod h1:IUgezN/MFpCDIlFezw3L8j83oeiIuYoj28Miwr/KUYo=
3 | github.com/FortAwesome/Font-Awesome v0.0.0-20221115183454-96cafbd73ec4 h1:xfr9SidRCMEh4A8fdkLhFPcHAVbrdv3Ua0Jp/nSmhhQ=
4 | github.com/FortAwesome/Font-Awesome v0.0.0-20221115183454-96cafbd73ec4/go.mod h1:IUgezN/MFpCDIlFezw3L8j83oeiIuYoj28Miwr/KUYo=
5 | github.com/FortAwesome/Font-Awesome v0.0.0-20230327165841-0698449d50f2 h1:Uv1z5EqCfmiK4IHUwT0m3h/u/WCk+kpRfxvAZhpC7Gc=
6 | github.com/FortAwesome/Font-Awesome v0.0.0-20230327165841-0698449d50f2/go.mod h1:IUgezN/MFpCDIlFezw3L8j83oeiIuYoj28Miwr/KUYo=
7 | github.com/google/docsy v0.5.1 h1:D/ZdFKiE29xM/gwPwQzmkyXhcbQGkReRS6aGrF7lnYk=
8 | github.com/google/docsy v0.5.1/go.mod h1:maoUAQU5H/d+FrZIB4xg1EVWAx7RyFMGSDJyWghm31E=
9 | github.com/google/docsy v0.6.0 h1:43bVF18t2JihAamelQjjGzx1vO2ljCilVrBgetCA8oI=
10 | github.com/google/docsy v0.6.0/go.mod h1:VKKLqD8PQ7AglJc98yBorATfW7GrNVsn0kGXVYF6G+M=
11 | github.com/google/docsy v0.7.0 h1:JaeZ0/KufX/BJ3SyATb/fmZa1DFI7o5d9KU+i6+lLJY=
12 | github.com/google/docsy v0.7.0/go.mod h1:5WhIFchr5BfH6agjcInhpLRz7U7map0bcmKSpcrg6BE=
13 | github.com/google/docsy/dependencies v0.5.1/go.mod h1:EDGc2znMbGUw0RW5kWwy2oGgLt0iVXBmoq4UOqstuNE=
14 | github.com/google/docsy/dependencies v0.6.0/go.mod h1:EDGc2znMbGUw0RW5kWwy2oGgLt0iVXBmoq4UOqstuNE=
15 | github.com/google/docsy/dependencies v0.7.0/go.mod h1:gihhs5gmgeO+wuoay4FwOzob+jYJVyQbNaQOh788lD4=
16 | github.com/twbs/bootstrap v4.6.2+incompatible h1:TDa+R51BTiy1wEHSYjmqDb8LxNl/zaEjAOpRE9Hwh/o=
17 | github.com/twbs/bootstrap v4.6.2+incompatible/go.mod h1:fZTSrkpSf0/HkL0IIJzvVspTt1r9zuf7XlZau8kpcY0=
18 | github.com/twbs/bootstrap v5.2.3+incompatible h1:lOmsJx587qfF7/gE7Vv4FxEofegyJlEACeVV+Mt7cgc=
19 | github.com/twbs/bootstrap v5.2.3+incompatible/go.mod h1:fZTSrkpSf0/HkL0IIJzvVspTt1r9zuf7XlZau8kpcY0=
20 |
--------------------------------------------------------------------------------
/src/content/en/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Xebia AI Toolbox
3 | ---
4 |
5 | {{< blocks/cover title="Xebia AI Toolbox" image_anchor="top" height="full" >}}
6 | At Xebia we believe AI can empower your work making you more productive in everything you do. This site helps you discover ways of empowering your daily workflows with AI in a responsible way.
13 | {{< blocks/link-down color="info" >}}
14 | {{< /blocks/cover >}}
15 |
16 |
17 | {{% blocks/lead color="primary" %}}
18 | We believe AI can help us if we use it in a responsible way. This means we are careful when working with AI tools regarding client data, privacy and security. We also believe in the power of open source and sharing knowledge. That is why we created this site to share our knowledge and help eachother in our daily work.
19 | {{% /blocks/lead %}}
20 |
21 |
22 | {{% blocks/section color="2" type="row" %}}
23 | {{% blocks/feature icon="fa-lightbulb" title="Start learning now!" url="docs" %}}
24 | Xebia AI Toolbox is a set of workflows created by and for Xebians to help eachother in their daily work becoming more productive using AI tools.
25 | {{% /blocks/feature %}}
26 |
27 |
28 | {{% blocks/feature icon="fab fa-github" title="Contributions welcome!" url="https://github.com/xebia/ai-toolbox" %}}
29 | We do a [Pull Request](https://github.com/xebia/ai-toolbox/pulls) contributions workflow on **GitHub**. New users are always welcome!
30 | {{% /blocks/feature %}}
31 |
32 |
33 | {{% blocks/feature icon="fab fa-slack" title="Discuss on Slack!" %}}
34 | Have questions or want to discuss something? Join us on **Slack** in the `#ai-toolbox` channel.
35 | {{% /blocks/feature %}}
36 |
37 |
38 | {{% /blocks/section %}}
39 |
40 |
41 | {{% blocks/section %}}
42 | *"AI will not replace you. A person using AI will."* - slackgpt
43 | {.h2 .text-center}
44 | {{% /blocks/section %}}
--------------------------------------------------------------------------------
/src/content/en/docs/AI Tools/slackgpt/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: SlackGPT
3 | weight: 4
4 | categories: [Tools]
5 | tags: [SlackGPT, ChatGPT]
6 | description: >
7 | ChatGPT alternative with responsible use of data in Xebia Slack
8 | ---
9 |
10 | |Responsible AI Checklist|
11 | |---|---|
12 | |Can be used with customer data| ℹ️ Ask Consent|
13 | |Can the output be used at customer| ✅ Yes|
14 | |Can be used with Xebia Internal secret data| ✅ Yes this tool is safe to use|
15 | |Can the output be used commercially by Xebia| ❓TODO VERIFY PROOF |
16 | |Is data being stored in the region you use it | Currently in Azure West Europe (Netherlands)|
17 | |Do we have a Xebia license / managed solution for this tool | ✅ Yes|
18 |
19 | ## What is SlackGPT
20 | SlackGPT was created at Xebia to offer an observable alternative to ChatGPT. The data that you send to SlackGPT is not stored anywhere, but does use the OpenAI GPT models. You can read more about the software [here][1].
21 |
22 | Because it uses a paid endpoint, our data will not be used to train ChatGPT or the GPT models.
23 |
24 | ## How to use SlackGPT
25 | SlackGPT is an app on the Xebia slack. You can chat with it directly. Every message will start a thread, which is the equivalent of ChatGPT's 'chat'.
26 |
27 | ## License / Costs
28 | The costs are covered by Xebia. Please use SlackGPT to your heart's content!
29 |
30 | ## Suitable to use with clients?
31 | It depends, if your client consents to their data being sent to the OpenAI endpoints, this is allowed. If you trigger the Microsoft content filters, the data might be viewed by humans at Microsoft.
32 |
33 | The guidance that SlackGPT will give you when starting a new thread:
34 |
35 | - The messages you send to SlackGPT are not stored or read by anyone within Xebia
36 | - We log usage metrics to track cost and get insights of how many people are using SlackGPT
37 | - Microsoft might review the messages if they suspect abuse of their API
38 |
39 | As a rule of thumb: don't put anything in here that you wouldn't put in Google Translate.
40 |
41 | [1]: https://xebia.com/blog/how-we-integrated-chatgpt-into-our-slack-enhancing-privacy-flexibility-and-collaboration/
42 |
--------------------------------------------------------------------------------
/src/content/en/docs/Responsible AI/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Responsible AI
3 | weight: 1
4 | description: >
5 | Thoughts and Considerations when using AI for your work
6 | ---
7 |
8 | ## Avoiding data leaks
9 | When using AI tools you often have to upload data to the cloud. This can be data from your clients or data from your company. You should always be careful when uploading client or company data to the cloud. You should always check if the data you are uploading is allowed and what will happen to this data.
10 |
11 | > ⚠️ Most free (and certain also paid) AI tools have policies that all data sent to it can be either stored and used for training their model or be used for other purposes. This means that if you upload client data to these tools you are violating the privacy of your client. This can have legal consequences for you and your company.
12 |
13 | In our [Tools](/docs/AI-Tools) section we describe which tools are safe to use, which aren't and what are the things you should consider when using them. It might not always be obvious which data a tool is using, for example IDE plugins that help you code often send parts of your solution to the cloud to help you code. This might be a problem when you are working on a client project that wants to keep their code secret.
14 |
15 | ## Always ask consent from your customer
16 | First rule for using any AI tool when working for your customer is to have consent from your customer in using these tools. Again the [Tools](/docs/AI-Tools) section has a description on which tools are safe and how to convince your customer to use them by providing proof to them of how the data is used/processed and what the risks are.
17 |
18 | ## You're The Pilot
19 | Whether you're using GitHub Copilot, or other AI tools, always remember, you're responsible for the end result. Review and reflect on the output of any AI tool.
20 |
21 | - Are there bugs in the generated code?
22 | - Is the generated material a blatant copy of a unique source?
23 | - Does the generated code create a security risk?
24 |
25 | ## Ethical use of AI
26 | Artificial Intelligence can be an incredibly strong tool. However, with great power, comes great responsibility. Whether you're building your own artificial intelligence, or using an AI tool, always be sure that it:
27 |
28 | - Is reliable and safe.
29 | - Does not discriminate against individuals or groups.
30 | - Does not use copyrighted material to generate results.
31 |
--------------------------------------------------------------------------------
/src/content/en/docs/Workflows/Creating Presentations from scratch/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Creating Presentations from scratch
3 | weight: 4
4 | categories: [Presentation]
5 | tags: [PowerPoint, SlackGPT]
6 | description: >
7 | How can SlackGPT help you with creating your presentations?
8 | ---
9 |
10 | |Responsible AI Checklist|
11 | |---|---|
12 | |Can be used with customer data input| ❌ No |
13 | |Can the output be used at customer | ❓ TODO check this|
14 | |Can be used with Xebia Internal secret data| ✅ yes|
15 | |Can the output be used commercially by Xebia | ✅ yes |
16 | |Is data being stored in the region you use it | ❌ No|
17 | |Do we have a Xebia license / managed solution for this tool | ✅ [SlackGPT](/docs/ai-tools/slackgpt)|
18 | |Tools used in workflow | [SlackGPT](/docs/ai-tools/slackgpt) |
19 |
20 | ## Introduction
21 | Creating a presentation can be hard. Especially the initial part. This workflow gets you going from nothing to a full presentation in powerpoint in a couple of minutes from which you can start adding styles and extra content.
22 |
23 | ## Step 1: Open the Visual Basic Editor
24 |
25 | Windows:
26 |
27 | Go to Developer[^enable-developer] > Visual Basic
28 |
29 | Mac:
30 |
31 | Go to Tools > Macro > Visual Basic Editor
32 |
33 | Next insert a Module: Insert > Module
34 |
35 | ## Step 2: Ask SlackGPT to do your homework
36 |
37 | Next up we are going to ask SlackGPT to generate a Visual Basic script that will generate a presentation for us. This is mostly Prompt engineering.
38 |
39 | Wether you know or you don't know your presentation out-line your prompt should probably start with something like this:
40 | ```
41 | Generate VBA script in PowerPoint that on execution generates slides.
42 | The Topic of the slide deck is: [The benefits of having ChatGPT do you work]
43 | The length of the slide deck is: [Minimum of 10 slides]
44 |
45 | Think like a C# developer that is trying to save his job by giving this presentation
46 | make sure to generate sub points on all the slides
47 | ```
48 |
49 | Run the script.
50 |
51 | ## Step 3: Style it
52 |
53 | All that is left is to style your slides, if you don't have a house-style design you can use the Integrated feature of PowerPoint "Designer" that should be located on the Home Ribbon.
54 |
55 | [^enable-developer]: If you don't see the Developer "Ribbon" follow these steps:
56 |
57 | - Go to File > Options
58 | - Customize Ribbon
59 | - Main Tabs > Developer
60 | - Ok
61 |
--------------------------------------------------------------------------------
/.github/workflows/hugo.yml:
--------------------------------------------------------------------------------
1 | # Sample workflow for building and deploying a Hugo site to GitHub Pages
2 | name: Deploy Hugo site to Pages
3 |
4 | on:
5 | # Runs on pushes targeting the default branch
6 | push:
7 | branches: ["main"]
8 |
9 | # Allows you to run this workflow manually from the Actions tab
10 | workflow_dispatch:
11 |
12 | # Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
13 | permissions:
14 | contents: read
15 | pages: write
16 | id-token: write
17 |
18 | # Allow only one concurrent deployment, skipping runs queued between the run in-progress and latest queued.
19 | # However, do NOT cancel in-progress runs as we want to allow these production deployments to complete.
20 | concurrency:
21 | group: "pages"
22 | cancel-in-progress: false
23 |
24 | # Default to bash
25 | defaults:
26 | run:
27 | shell: bash
28 |
29 | jobs:
30 | # Build job
31 | build:
32 | runs-on: ubuntu-latest
33 | env:
34 | HUGO_VERSION: 0.113.0
35 | steps:
36 | - name: Install Hugo CLI
37 | run: |
38 | wget -O ${{ runner.temp }}/hugo.deb https://github.com/gohugoio/hugo/releases/download/v${HUGO_VERSION}/hugo_extended_${HUGO_VERSION}_linux-amd64.deb \
39 | && sudo dpkg -i ${{ runner.temp }}/hugo.deb
40 | - name: Install Dart Sass Embedded
41 | run: sudo snap install dart-sass-embedded
42 | - name: Checkout
43 | uses: actions/checkout@v4
44 | with:
45 | submodules: recursive
46 | - name: Setup Pages
47 | id: pages
48 | uses: actions/configure-pages@v3
49 | - name: Install Node.js dependencies
50 | working-directory: ./src
51 | run: "[[ -f package-lock.json || -f npm-shrinkwrap.json ]] && npm ci || true"
52 | - name: Build with Hugo
53 | working-directory: ./src
54 | env:
55 | # For maximum backward compatibility with Hugo modules
56 | HUGO_ENVIRONMENT: production
57 | HUGO_ENV: production
58 | run: |
59 | hugo \
60 | --minify \
61 | --baseURL "${{ steps.pages.outputs.base_url }}/"
62 | - name: Upload artifact
63 | uses: actions/upload-pages-artifact@v1
64 | with:
65 | path: ./src/public
66 |
67 | # Deployment job
68 | deploy:
69 | environment:
70 | name: github-pages
71 | url: ${{ steps.deployment.outputs.page_url }}
72 | runs-on: ubuntu-latest
73 | needs: build
74 | steps:
75 | - name: Deploy to GitHub Pages
76 | id: deployment
77 | uses: actions/deploy-pages@v2
78 |
--------------------------------------------------------------------------------
/src/content/en/docs/Workflows/Creating Photorealistic Images/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Creating Photorealistic Images
3 | weight: 4
4 | categories: [Images, Presentation]
5 | tags: [ChatGPT, SlackGPT, Midjourney]
6 | description: >
7 | How can you combine ChatGPT and Midjourney to create photorealistic images?
8 | ---
9 |
10 | |Responsible AI Checklist|
11 | |---|---|
12 | |Can be used with customer data input| ❌ No|
13 | |Can output be used at customer | ✅ Yes|
14 | |Can be used with Xebia Internal secret data| ❌ No|
15 | |Can output be used commercially by Xebia | ✅ Yes|
16 | |Is data being stored in the region you use it | ❌ No|
17 | |Do we have a Xebia license / managed solution for this workflow | ❌ No|
18 | |Tools used in workflow | [Midjourney](/docs/ai-tools/midjourney), [ChatGPT](/docs/ai-tools/chatgpt) / [SlackGPT](/docs/ai-tools/slackgpt) |
19 |
20 | ## Introduction
21 | Do you want to create photorealistic images with Midjourney, but get stuck when you need to specify your prompt? You can teach ChatGPT/SlackGPT to generate prompts for you. The generated prompt includes style, camera type, and even camera settings. This document describes the process, but the initial script can be changed to create other styles of images.
22 |
23 | ## Important: Credit
24 | The steps described on this page were designed by the YouTube channel [AI Foundations][1]. The author created a [video][2] on this subject that greatly improves the Midjourney experience.
25 |
26 | ## Steps
27 | ### Step 1: Add the introduction script to ChatGPT
28 | Download [the script][3] on the original author's website. You can paste this script into ChatGPT, or even SlackGPT. After you've sent the script, the AI assistant will respond with a resounding "YES".
29 |
30 | ### Step 2: Tell the assistant about your subject
31 | What kind of picture do you need? I wanted to generate a 'A photo of a rubber duck sitting on a monitor'. SlackGPT responded with:
32 | > Realistic photograph of a cheerful yellow rubber duck, playfully perched on top of a computer monitor in a cozy office setting, soft ambient lighting, blurred background showcasing office desk and accessories, Nikon D850, 50mm lens, f/1.8, natural light --ar 3:2 --v 5 --q 1
33 |
34 | ### Step 3: Give the prompt to Midjourney
35 | To get started with Midjourney, [start here][5]. Currently, Midjourney is only accessible through Discord.
36 | Copy the prompt, and prefix it with `/imagine`. Paste this into Midjourney on Discord and it will generate options for you. Here's the result:
37 | ![duckies][4]
38 |
39 | Note: All Midjourney prompts and results are publicly visible, unless you have a premium subscription with Stealth Mode enabled.
40 |
41 | ## Licenses needed
42 | ### SlackGPT
43 | You can use [SlackGPT][6] on the Xebia Slack. It's a great alternative for ChatGPT.
44 |
45 | ### Midjourney
46 | You need a personal license. Find out more about Midjourney [here][7].
47 |
48 |
49 | [1]: https://www.youtube.com/@ai-foundations
50 | [2]: https://www.youtube.com/watch?v=EJez32MtfWU
51 | [3]: https://aiwealth.vip/midjourney-training-sheet/
52 | [4]: ./duckies.png
53 | [5]: https://docs.midjourney.com/docs/quick-start
54 | [6]: /docs/ai-tools/slackgpt
55 | [7]: /docs/ai-tools/midjourney
56 |
--------------------------------------------------------------------------------
/src/content/en/docs/Workflows/Creating a annotated speaker deck for training/_index.md:
--------------------------------------------------------------------------------
1 | ---
2 | title: Creating an Annotated speaker deck for training purposestions from scratch
3 | weight: 4
4 | categories: [Presentation]
5 | tags: [PowerPoint, SlackGPT, WhisperAI]
6 | description: >
7 | How can you create a PowerPoint deck with speaker notes so your presentation or training can be redelivered?
8 | ---
9 |
10 | |Responsible AI Checklist|
11 | |---|---|
12 | |Can be used with customer data input| ❌ No |
13 | |Can the output be used at customer | ❓ TODO check this|
14 | |Can be used with Xebia Internal secret data| ✅ yes|
15 | |Can the output be used commercially by Xebia | ✅ yes |
16 | |Is data being stored in the region you use it | ❌ No|
17 | |Do we have a Xebia license / managed solution for this tool | ✅ [SlackGPT](/docs/ai-tools/slackgpt)|
18 | |Tools used in workflow | [SlackGPT](/docs/ai-tools/slackgpt), [Whisper AI](/docs/ai-tools/whisper) |
19 |
20 | ## Introduction
21 | When you create a PowerPoint deck that can be used in a training or other sort of multi-delivery, it is hard to capture the story you want to convey in slides only. What do ou tell at every slide, what are the talking points and what story line do you want to follow. When you deliver the presentation yourself, this is doable, but what if someone else needs to use your deck? Or you deliver a training deck to a customer. For example, a customer wants a deck they can use for training their people. Then you want to annotate the slides with speaker notes, the story you want to tell so that people can read it, and the talking points so people can create their own story without losing the storyline.
22 |
23 | No better way to capture this by recording the presentation right? Right! Bu that does not really scale well when preparing or delivering. So you want your story in the speaker notes.. But that is a lot of work! Yes indeed. And that is where this workflow comes in
24 |
25 | ## Step 1: Deliver your presentation to your recording device
26 | Deliver the presentation. On Windows, use the audio recorder and deliver your story as you would do it in a training or presentation. Start every new slide with