Skip to main content

Configuring AI models

How to configure AI provider API keys and manage AI model settings for Storydesk's writing and generation features.

Cuez Admin is where you connect AI providers to your organization so that Storydesk's AI writing, summarization, and story discovery features work for your team. You configure two types of models: an LLM for chat and script assistance, and an Embedding model for story linking and similarity search. Organization owners and group admins can both configure AI models.

What AI models enable in Storydesk

Once configured, AI models power the following features in Storydesk:

  • AI commands in the script editor, by typing /, such as writing a brief, fact-checking, summarizing, and more

  • AI drafting in the ‘Social’ and ‘Web’ tabs

  • The Agent

  • Story linking and similarity search

  • AI story discovery (requires an interest profile configured per group, see Setting up news feeds)

If no AI model is configured for your organization, these features do not appear in your Storydesk.


​

Prerequisites

AI features must be enabled for your organization by Cuez. If AI Models is not visible in the sidebar, contact [email protected].
​

Configuring an LLM

  1. Navigate to AI Models in the left sidebar.

  2. In the ‘LLM’ card, click ‘+ Configure LLM’.

  3. Select your AI provider and enter the API key for that provider.

  4. Complete any additional fields shown for your selected provider.

  5. Save the configuration.

Once saved, the LLM card shows the configured model and provider.


​

Configuring an Embedding model

  1. Navigate to AI Models in the left sidebar.

  2. In the ‘Embedding’ card, click ‘+ Configure Embedding’.

  3. Select your provider and enter the API key.

  4. Save the configuration.

The Embedding model is used for semantic story search and feed relevance scoring. It operates in the background and does not require any action from Storydesk users.


​

Viewing token usage

The 'Token Usage' section at the bottom of the AI Models page shows how your organization is consuming AI tokens over time. Use the time period dropdown to select the current month or a previous period.
​

Switch between views using the tabs:

  • By Feature - breaks down usage across language model features (Script AI Assistant, Fact Checking, Story Finder, Social Media AI, Web Publishing AI, Brief vs Rundown Analysis) and embedding features (Story Embeddings, Feed Embeddings)

  • By Hour - shows usage distribution across hours of the day

  • By Weekday - shows usage distribution across days of the week
    ​

Verifying the setup

After configuring both models, log in to Storydesk, open a story, and type / in the script editor. The AI command menu should appear with options such as 'Write story' and 'Fact check'. If the menu does not appear, check that both models are saved on the AI Models page and that AI features are enabled for your organization.

Did this answer your question?