We're excited to announce AI Triggers, a powerful new feature that brings automation to your development workflow. Plus, Pro users now have the ability to choose their AI model for each trigger.
What Are AI Triggers?
AI Triggers let you define automated actions that fire when specific development summaries arrive. Whether it's generating tweet drafts from marketing summaries, creating documentation from architect notes, or drafting release notes from product updates - triggers handle it automatically.
How Triggers Work
- Configure a trigger with a name, event filter, and prompt template
- Set the execution mode - auto-execute or require approval
- When a matching summary arrives, the trigger fires and generates output
Event Filters
Triggers can fire on specific summary types:
- Marketing summaries - User-facing progress and business value
- Architect summaries - Technical details and implementation decisions
- Product summaries - Feature status and user story progress
- Bugfix summaries - Bug descriptions and fix details
- Documentation summaries - Content updates
- Any summary - Fire on all summary types
Template Variables
Your prompt templates can use these variables:
{{marketing_summary}}- The marketing summary content{{architect_summary}}- The architect summary content{{product_summary}}- The product summary content{{workspace_name}}- The name of the workspace{{branch}}- The current branch name
Example Triggers
Tweet Generator
Name: Tweet from Marketing
Trigger when: Marketing summary received
Prompt: Write a tweet about: {{marketing_summary}}
Release Notes Drafter
Name: Release Notes
Trigger when: Product summary received
Prompt: Draft release notes based on: {{product_summary}}
Include the key features and any breaking changes.
Technical Documentation
Name: Architecture Docs
Trigger when: Architect summary received
Prompt: Update the architecture documentation with:
{{architect_summary}}
Focus on the design decisions and patterns used.
AI Model Selection (Pro)
Pro users can now choose which AI model powers each trigger:
Haiku (Fast)
Perfect for quick, simple tasks like tweet generation or brief summaries. Fastest response times and lowest cost.
Sonnet (Recommended)
The default choice. Great balance of speed, quality, and capability. Ideal for most trigger use cases.
Opus (Most Capable)
For complex tasks requiring deep reasoning. Best for detailed documentation, technical analysis, or nuanced content generation.
Free and Basic users default to Sonnet for all triggers, which provides excellent results for most use cases.
Execution Modes
Auto-Execute
The trigger runs immediately when a matching summary arrives. Output appears in your triggers panel automatically.
Require Approval
The trigger creates a pending output that you can review before executing. Perfect for important content that needs human oversight.
Outputs Panel
All trigger outputs are displayed in a dedicated panel where you can:
- View results - See the AI-generated content
- Copy to clipboard - One-click copy for easy sharing
- Track status - Pending, Running, Completed, or Failed
- Approve or cancel - For triggers requiring approval
Getting Started
- Open the Triggers Panel from the sidebar
- Click Add Trigger
- Configure your trigger settings
- Enable the trigger and watch it work
Triggers persist across sessions, so your automation setup is always ready.
Pro Features
AI Model Selection is available on the Pro plan ($189/month), which also includes:
- 5 devices
- Unlimited workspaces
- Activity Insights (AI-powered)
- Premium support
Upgrade to Pro to unlock AI Model Selection for triggers.
What's Next
We're continuing to expand the trigger system with:
- More action types - Direct posting to platforms, creating issues, updating documentation
- Conditional logic - Triggers that only fire under certain conditions
- Chaining - Triggers that can invoke other triggers
Have ideas for trigger improvements? Let us know on GitHub.