3 months ago

Feature: add Default AI Provider in AI Hub Settings, used by AI Intent Recognition and AI Task

Default AI Provider Setting in AI Hub
A new Default AI Provider setting is now available in AI Hub → Settings — allowing you to configure the global default AI provider and model used by AI Intent Recognition and AI Tasks across all flows in the workspace.
What's New
  • 🤖 New setting: Default AI Provider — found at AI Hub → Settings (gear icon in the AI Hub submenu)
  • 🔌 Select from any supported AI provider: OpenAI, Anthropic (Claude), xAI (Grok), Groq, and more
  • 🧩 Set a specific Model within the chosen provider (e.g., gpt-4.1 for OpenAI)
  • 📋 A Default model hint is shown below the model field (e.g., "Default model: gpt-4o-mini") to indicate the fallback when no model is specified
  • 💾 Click Save to apply — takes effect immediately across all AI features that use the default provider
How to Use
  1. Go to AI Hub in the left sidebar
  2. Click Settings (gear icon, bottom of the AI Hub submenu)
  3. Under Default AI Provider, select your preferred provider from the dropdown (e.g., OpenAI)
  4. Enter a specific Model name if desired (e.g., gpt-4.1) — leave blank to use the provider default
  5. Click Save
Full documentation: AI Hub Settings
Where It's Used
The Default AI Provider setting is the fallback provider used when no specific provider has been configured at the individual feature level:
  • AI Intent Recognition — uses the default provider when no override is set on the intent node
  • AI Tasks — uses the default provider when no specific provider is set on the AI Task
  • Additional AI features may inherit this default in future updates
Why It Matters
  • Centralises AI provider configuration — set it once in AI Hub Settings and all AI features use it consistently
  • Eliminates the need to configure a provider on every individual AI Intent node or AI Task
  • Makes it easy to switch all AI features to a new provider or model in a single change
  • Pairs with the Partner AI Key settings (OpenAI, xAI, Groq) for a fully managed AI stack across partner workspaces