We're excited to introduce AI Discoverability, a new feature that makes your public wiki content accessible to AI assistants and Large Language Models (LLMs) through the standardized llms.txt format.
What is llms.txt?
Similar to how robots.txt helps search engines understand your site, llms.txt helps AI assistants understand and reference your public documentation. When enabled, your wiki automatically generates a llms.txt file at /t/[your-workspace-slug]/llms.txt.
Key Features
Opt-in by Default
AI discoverability is disabled by default. You have full control over whether your public content is exposed to AI tools.
Public Content Only
Only sections and pages marked as public are included in the llms.txt file. Private content is never exposed.
Metadata Only
The llms.txt file contains only page titles, summaries, and section descriptions - not full page content. This keeps the file lightweight while providing enough context for AI tools.
Custom Descriptions
Add a custom description to help AI tools understand the purpose and context of your documentation.
Performance Optimized
The llms.txt file is cached for 1 hour with stale-while-revalidate support for up to 24 hours, ensuring fast responses even for large wikis.
How to Enable
- Navigate to Settings in your workspace
- Find the AI Discoverability section
- Toggle Enable llms.txt for public wiki
- (Optional) Add a custom description
- Click Save Settings
Use Cases
- Developer Documentation: Make API docs discoverable to AI coding assistants
- Product Documentation: Help AI tools answer questions about your product
- Knowledge Bases: Enable AI assistants to reference your public knowledge base
- Educational Content: Make tutorials accessible to AI-powered learning tools
Privacy & Security
- Only workspace owners and admins can enable/disable this feature
- Only public sections and pages are included
- Full page content is never exposed
- The endpoint returns 404 when disabled
This feature is available now for all team accounts. Learn more in the AI Discoverability documentation.