AyselAI CLI
Write lessons on your computer, try them exactly like AyselAI runs them, upload them as drafts and send courses for review — from the terminal or CI.
It can do exactly what your AI app can do through AyselAI's connection, nothing more: drafts only, publishing happens in AyselAI.
Download (Node 20 or newer)
One file. The second line checks it is exactly the file we published.
curl -fsSL https://aysel.ai/cli/ayselai.mjs -o ayselai.mjs curl -fsSL https://aysel.ai/cli/ayselai.mjs.sha256 | shasum -a 256 -c - node ayselai.mjs --version
To type just ayselai (macOS, Linux):
mkdir -p ~/.local/bin && mv ayselai.mjs ~/.local/bin/ayselai && chmod +x ~/.local/bin/ayselai
npm (npm install -g) comes later.
Sign in
Your browser opens the AyselAI permission screen; allow "AyselAI CLI". You can remove it any time under Connected apps.
ayselai login ayselai courses list
No browser on that machine (SSH)? ayselai login --manual shows a link to open anywhere and asks for the code.
Your first lesson
ayselai lesson init my-lesson ayselai lesson preview my-lesson ayselai lesson push my-lesson --section <section-id>
Tab completion, checks, updates
doctor checks your setup; update tells you when a newer version is out.
ayselai completion bash > ~/.local/share/bash-completion/completions/ayselai ayselai completion fish > ~/.config/fish/completions/ayselai.fish ayselai doctor ayselai update
Automatic uploads (CI)
Create an upload key for one course and put it in your CI as AYSELAI_TOKEN. The first upload of a new lesson is done once from your computer (it remembers the lesson in .ayselai-lesson.json; commit that file).
Anyone who has the key can upload drafts to that course until it expires or you revoke it. Put it only in your CI's secret settings (never in code, a chat or an e-mail). We show it once.
GitHub Actions example
name: AyselAI lessons on: push: branches: [main] pull_request: jobs: lessons: runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - uses: actions/setup-node@v4 with: node-version: 22 - name: Download the AyselAI CLI (checksum verified) run: | curl -fsSL https://aysel.ai/cli/ayselai.mjs -o ayselai.mjs curl -fsSL https://aysel.ai/cli/ayselai.mjs.sha256 | shasum -a 256 -c - - name: Check the lesson run: node ayselai.mjs lesson validate lessons/my-lesson - name: Upload a new draft version if: github.event_name == 'push' env: AYSELAI_TOKEN: ${{ secrets.AYSELAI_TOKEN }} AYSELAI_SERVER: https://aysel.ai run: node ayselai.mjs lesson push lessons/my-lesson
Your sign-in stays in your system keychain (or a file only you can read) and is never printed. The CLI never sees learners' data.