Weaver Interface
Weaver’s interface combines a chat panel and visual Canvas into a single workspace. You describe your goals in natural language, and Weaver begins mapping the workflow while you watch it come to life.
If you want to see practical examples of prompts, browse the Prompt Library. To follow a step-by-step build, try the Weaver Quickstart.
Main Panels
Weaver’s interface has three main areas that work together:
Panel | What You’ll See |
---|---|
Chat Panel | Where you describe your goals, view Weaver’s “thinking aloud,” and respond to follow-up questions. Logs and test results also appear here. |
Planned Workflows & Skill Suggestions | A draft plan of the steps Weaver thinks the workflow should include, along with recommended components for each. |
Component Canvas | A live diagram of your workflow. Components appear and connect here automatically, and you can edit them directly at any time. |
How Workflows Are Planned
When you provide a goal, Weaver drafts a workflow before building it. For a blog-writing request, the plan could include:
- Generate Outline (titles and headings)
- Research Supporting Material
- Write Full Draft
- Optional: Publish Blog
To refine the plan, Weaver often asks questions such as:
- “Do you want technical, lifestyle, or marketing content?”
- “How long should the blog usually be?”
This back-and-forth ensures the workflow reflects your actual needs instead of assumptions.
Working with Components
Once the Canvas begins to fill with components, you can interact with them directly:
- Click a block to open its inspector.
- Ctrl/Cmd + click to select multiple blocks for editing or re-running.
- Shift + click empty space to clear your selection.
Weaver will automatically suggest components, but you can also add your own. The Component Library lists all available blocks, such as API connectors, content generators, and transformers.
Using Diagrams as Inputs
You can drag and drop PNG, JPG, or SVG files into the chat. Weaver interprets these diagrams as a starting point by:
- Extracting text from the image.
- Reading arrows as potential data flows.
- Suggesting components that match what it sees.
Contextual Actions
Weaver provides contextual actions that appear when you hover or click on components:
Action | When It Appears | What It Does |
---|---|---|
Fix with AI | When a component shows red or orange | Opens a pre-filled repair prompt with error context so Weaver can fix the issue. |
Explain Component | Any time | Gives a plain-language explanation of what the block does, including inputs and outputs. |
Trash | On hover | Deletes the component and reroutes connections when possible. |
Next Steps
The interface is only one part of working with Weaver. To keep learning:
- Explore the Prompt Library for practical examples across use cases.
- Walk through the Weaver Quickstart to build your first complete agent.
- Review available components in the Component Library to understand what Weaver uses when wiring workflows.
- Refer to Debugging Agents whenever you run into issues.