Canoryn

Appearance

Quick Start: Build a "Zen Master" ​

In this guide, you will build an agent that finds ancient wisdom and turns it into deep philosophical musings.

The Mission: Fetch a Zen quote from GitHub's API, have AI reflect on it, and speak it aloud.

You will learn how to:

  1. Fetch Data from an API (HTTP Request).
  2. Process Data with an AI model.
  3. Perform Actions (Text-to-Speech).

Step 1: Create a New Agent ​

  1. Open the Architect (Window > Architect).
  2. In the Sidebar, click the + button (or right-click "My Projects" -> "New Agent").
  3. Name your agent "Zen Master".

Screenshot coming soon: Screenshot: Empty Canvas and New Agent Dialog

Step 2: Fetch Wisdom ​

We'll use GitHub's Zen API to get a random design principle in plain text. It's fast, reliable, and free.

  1. Right-click on the canvas to open the Node Palette.
  2. Search for "HTTP" and select HTTP Request (System).
  3. Click the new node to select it. In the Inspector (Right Panel), configure:
    • URL: https://api.github.com/zen
    • Method: GET.

Tip: Press Play in the toolbar now. Check the Console (Bottom Panel) to see the output (e.g., "Favor focus over features.").

Screenshot coming soon: Screenshot: HTTP Request Node configuration

Step 3: Add the AI Brain ​

Now let's expand on this thought.

  1. Add a Simple AI node (Logic).
  2. Drag a wire from the HTTP Request node's Result port (Green/String).
  3. Connect it to the Simple AI node's Context port.
  4. Set the System Instruction in the Inspector:
    text
    You are a wise master.
The context contains a design principle.
Reflect on it and speak a short, inspiring message about life.
Keep it under 2 sentences.

Screenshot coming soon: Screenshot: Simple AI Node wired to HTTP Request

Step 4: Give it a Voice ​

Finally, let's hear the wisdom.

  1. Add a Speak Text node (General).
  2. Connect the Simple AI node's Result port to the Speak Text node's Text input.
  3. (Optional) Select a Personality in the Inspector (e.g., "Narrator").

Screenshot coming soon: Screenshot: Speak Text Node wired to Simple AI

Step 5: Run It! ​

  1. Turn up your volume 🔊.
  2. Press the Play button (▶️).
  3. Watch the "Marching Ants" flow through the graph.
  4. Your computer will speak a profound truth!

Why this works ​

You just built a Linear Pipeline: [Data Source] -> [Intelligence] -> [Action]

This simpler GitHub API returns plain text, so you didn't need any complex parsing nodes. To handle complex APIs (like Notion or Linear), you would use the JSON Parser node in between.