Getting Started

Favorites and Collections Guide

Organize your AI models for quick access with favorites and collections. This guide covers how to save models you love and create custom groups for different workflows.

Favorites

What Are Favorites?

Favorites are models you've starred for quick access:

  • Appear in dedicated Favorites tab
  • Show first in model selectors
  • Sync across devices
  • Easy one-click access

Adding to Favorites

From the Models Page:

  1. Find the model you want to favorite
  2. Click the star icon on the model card
  3. The star fills in to confirm
  4. Model appears in your Favorites tab

From the Chat Model Selector:

  1. Open the model selector
  2. Hover over any model
  3. Click the star icon
  4. Model is added to favorites

Viewing Favorites

Access your favorites in multiple places:

Models Page:

  1. Navigate to Models
  2. Click the Favorites tab
  3. All starred models appear

Chat Model Selector:

  1. Open the model selector
  2. Click the Favorites tab
  3. Quick access to starred models

Removing from Favorites

Click the filled star icon to remove:

  1. Find the favorited model
  2. Click the star (now filled)
  3. Model is removed from favorites
  4. Star returns to outline state

Favorites Sync

Your favorites automatically sync:

  • Saved to your account
  • Available on all devices
  • Persist across sessions
  • Backed up to localStorage for offline access

Collections

What Are Collections?

Collections are custom model groups with intelligent routing:

  • Group related models together
  • Apply routing strategies
  • Use as a single "model" in chat
  • Perfect for testing and failover

Creating a Collection

Step 1: Navigate to Collections

  1. Go to Models page
  2. Click the Collections tab
  3. Click Manage Collections or navigate to /model-collections

Step 2: Create New Collection

  1. Click Create Collection
  2. Enter a name (used as the collection ID)
  3. Add a display name and description
  4. Select your routing strategy

Step 3: Add Models

  1. Click Add Models in your collection
  2. Search or browse available models
  3. Click models to add them
  4. Set priority order if using priority routing

Collection Properties

Each collection has:

Property Description
Name Unique identifier (collection/name)
Display Name Human-friendly name
Description Optional details
Strategy How to select models
Members Models in the collection

Routing Strategies

Collections support different routing strategies:

Random

  • Randomly selects a model per request
  • Good for load distribution
  • Tests model variety

Round Robin

  • Cycles through models in order
  • Even distribution
  • Predictable rotation

Priority

  • Uses first available model
  • Falls back to next in order
  • Great for failover

Least Used

  • Selects least recently used
  • Balances over time
  • Prevents overuse

Using Collections in Chat

Use a collection like any model:

Method 1: From Collections Tab

  1. Go to Models > Collections
  2. Find your collection
  3. Click Use in Chat
  4. Chat opens with collection selected

Method 2: Direct Selection

  1. Open chat model selector
  2. Go to Collections tab
  3. Select your collection
  4. Start chatting

Method 3: By ID The collection ID format is:

collection/your-collection-name

Collection Examples

Quality Comparison:

Name: quality-test
Strategy: Random
Models:
- gpt-4o
- claude-3-5-sonnet
- gemini-1.5-pro

Use case: Compare quality across providers

Cost-Optimized Fallback:

Name: budget-first
Strategy: Priority
Models:
1. gpt-4o-mini (first choice)
2. claude-3-5-haiku (fallback)
3. gpt-4o (premium fallback)

Use case: Use cheapest first, fall back as needed

Provider Diversity:

Name: multi-provider
Strategy: Round Robin
Models:
- openai/gpt-4o
- anthropic/claude-3-5-sonnet
- google/gemini-1.5-pro

Use case: Distribute across providers

Managing Model Preferences

Recent Models

Recent models are tracked automatically:

  • Last used models appear first
  • Updated with each chat session
  • Quick access in model selector
  • Cleared when you want fresh start

Model Notes

Add personal notes to models (when available):

  • Record use cases
  • Note strengths and weaknesses
  • Track observations
  • Share with team

Default Model

Set a default model for new chats:

  1. Favorite your preferred model
  2. Use it in chat
  3. It becomes the default for new sessions

Or explicitly set in preferences:

  1. Go to Settings
  2. Find Chat Preferences
  3. Select default model

Ratings and Reviews

Rating Models

Provide feedback on model quality:

Leaving a Rating:

  1. Open model details
  2. Click the rating section
  3. Select star rating (1-5)
  4. Optionally add a comment
  5. Submit your rating

Rating Criteria

Consider these factors:

Factor Description
Quality Response accuracy and helpfulness
Speed Response latency
Value Quality relative to price
Reliability Consistency of results

Viewing Ratings

See community ratings:

  1. Open model details
  2. View aggregate rating
  3. Read individual reviews
  4. Filter by rating level

Rating Guidelines

When rating models:

  • Be objective and specific
  • Consider your use case
  • Update ratings as models improve
  • Helpful reviews get highlighted

Organization Tips

Naming Conventions

Use consistent naming for collections:

By Purpose:

writing-assistants
code-helpers
research-tools

By Quality Tier:

premium-models
budget-models
experimental

By Project:

project-alpha-models
marketing-content
technical-docs

Collection Organization

Tips for effective collections:

  1. Keep focused - Each collection has a purpose
  2. Limit size - 3-5 models per collection is ideal
  3. Update regularly - Remove outdated models
  4. Document purpose - Use descriptions effectively

Favorites Strategy

Organize favorites by:

  1. Daily drivers - Models you use constantly
  2. Specialists - Best for specific tasks
  3. Experiments - New models to try
  4. Fallbacks - Reliable alternatives

Sharing and Collaboration

Sharing Collections

Collections can be shared with your organization:

  1. Create a collection
  2. Set visibility to organization
  3. Team members can view and use
  4. Useful for standardizing model use

Team Favorites

While personal favorites are private:

  • Share recommendations in team chat
  • Create shared collections for teams
  • Document favorite models in wikis

Troubleshooting

Favorites Not Syncing

If favorites don't appear on other devices:

  1. Ensure you're logged in
  2. Check internet connection
  3. Wait a moment for sync
  4. Refresh the page

Collection Not Working

If collection routing fails:

  1. Verify all models are available
  2. Check model permissions
  3. Try individual models
  4. Review collection settings

Missing Models

If expected models don't appear:

  1. Check your organization access
  2. Verify connection status
  3. Search by model ID
  4. Contact admin if needed

Best Practices

For Favorites

  1. Be selective - Too many favorites defeats the purpose
  2. Review regularly - Remove models you no longer use
  3. Consider task fit - Favorite task-appropriate models
  4. Test before favoriting - Ensure models work for you

For Collections

  1. Clear purpose - Each collection serves a goal
  2. Test strategies - Find what works for your workflow
  3. Monitor performance - Check which models get used
  4. Iterate - Adjust based on results

Next Steps