#AI Agents

AI Agents are Hygraph’s intelligent automation assistants designed to handle repetitive content workflows, such as translation, summarization, and SEO optimization. Agents help drive predefined content workflows autonomously. They help teams scale content operations while ensuring quality, consistency, and editorial control.

With Agents, teams can automate routine tasks through workflows, maintain a consistent brand voice, all without needing developer intervention.

#How it works

Agents are powered by integrated large language models (LLMs) that understand your project’s schema and context. They interact with your content safely and intelligently, applying automation to specific fields or entries based on defined actions.

Each agent is specialized for a distinct purpose, such as translation, summarization, or SEO, and works according to the configuration you define. Once set up, agents can run automatically or be triggered manually, ensuring your workflows stay efficient and flexible.

#Model information

Hygraph uses the latest Claude models to power AI tasks such as SEO, content summarization, and translation. These models are selected for their quality, accuracy, and efficiency.

We are continuously testing new models to improve performance and reliability. Model versions may change over time as newer releases become available, and updates are applied automatically to maintain optimal results.

#Benefits

Agents help teams automate common editorial workflows and maintain consistency at scale.

• Save time by automating repetitive tasks like generating SEO metadata or translating fields.
• Maintain tone and structure consistency across multiple locales.
• Automatically summarize or condense long-form content into concise, context-aware summaries.
• Operate securely within your content model by respecting existing roles, permissions, and content governance.

#Get started

#Availability

• AI Agents is an enterprise feature and is currently in Early Access. Contact our sales team for access.

#Permissions

Agents respect user roles and content permissions. Only users with permissions to set up Content Workflows can set up an agent.

#Set up Agents

1. Navigate to Project Settings → AI Agents configuration.
2. Click Add Agent.
3. Choose the agent type:
   • SEO expert
   • Translator agent
   • Content summarizer
4. Configure the agent’s behavior, and click Add agent.

#SEO expert

Use the SEO expert to generate an SEO Report with feedback on how well content is optimized for search. Follow these steps to configure the SEO expert.

1. Navigate to Project Settings → AI Agents configuration.
2. Click Add Agent, and select Set up next to the SEO expert.
3. Provide a name to the agent, and optionally provide a description.
4. Under Number of seats, specify the number of agent instances that can run concurrently for the tasks.
5. Select the Workflow where the agent should operate. The agent is triggered only for those entries that move through this workflow.
6. Choose the Workflow step where the agent should run. This ensures the agent runs at the right stage in your editorial process. You can choose an existing workflow step, or add a step to the workflow. If you need to add a step to the workflow, select Add a new AI Agent steps, and do the following:
   1. Under Execute after step, choose the step after which you want to add this step in the workflow.
   2. Provide a Step API ID and Step Display Name.
7. Select the Content model that contains the fields to provide feedback for SEO optimization.
8. Under Fields to optimize, select one or more fields that the SEO agent should analyze and provide feedback on.
9. Select Add agent.

Once added, the SEO expert runs automatically during the configured workflow step.

#Translation agent

Use the Translation Agent to automatically translate content between locales while preserving tone, structure, and meaning. It helps teams scale multilingual content creation and ensures consistent messaging across all supported languages. Follow these steps to configure the Translation agent.

1. Navigate to Project Settings → AI Agents configuration.
2. Click Add Agent, and select Set up next to the Translation agent.
3. Provide a name for the agent, and optionally provide a description.
4. Under Number of seats, specify the number of agent instances that can run concurrently for translation tasks.
5. Select the Workflow where the agent should operate. The agent is triggered only for entries that move through this workflow.
6. Choose the Workflow step where the agent should run. This ensures the agent runs at the correct stage in your localization process. You can choose an existing workflow step or add a new one. If you need to add a new workflow step, select Add a new AI Agent step, and do the following:
   1. Under Execute after step, choose the step after which this translation step should run.
   2. Provide a Step API ID and a Step Display Name.
7. Under Select languages, specify the locales you want to translate into. You can select one or multiple target locales depending on your project’s localization setup.
8. Under Custom instructions, provide optional translation guidance to help maintain tone and context. For example, “Use formal tone for business audiences” or “Keep product names untranslated”.
9. Select Add agent.

Once added, the Translation Agent automatically runs during the configured workflow step. It translates content into the target locales while preserving structure, tone, and meaning according to your custom instructions.

#Content summarizer

Use the Content summarizer to automatically generate concise summaries of long-form content. It helps teams create consistent excerpts, overviews, or meta descriptions across entries, saving time while maintaining quality and tone. Follow these steps to configure the Summarization agent.

1. Navigate to Project Settings → AI Agents configuration.
2. Click Add Agent, and select Set up next to the Content summarizer.
3. Provide a name for the agent, and optionally provide a description.
4. Under Number of seats, specify the number of agent instances that can run concurrently for summarization tasks.
5. Select the Workflow where the agent should operate. The agent is triggered only for entries that move through this workflow.
6. Choose the Workflow step where the agent should run. This ensures the agent runs at the correct stage in your editorial process. You can choose an existing workflow step or add a new one. If you need to add a new workflow step, select Add a new AI Agent step, and do the following:
   1. Under Execute after step, choose the step after which this summarization step should run.
   2. Provide a Step API ID and a Step Display Name.
7. Under Tone, choose the style or tone that the generated summaries should follow.
8. Under Custom instructions, provide optional guidance to help control summary style, structure, or length. For example, “Summarize in 2–3 sentences,” or “Focus only on key benefits.”
9. Select Add agent.

Once added, the Summarization Agent automatically runs during the configured workflow step. It generates short, context-aware summaries for the selected entries while maintaining the specified tone and following your custom instructions.

#Manage Agents

After creating agents, you can manage and monitor their configuration and performance directly from your project settings. This includes editing details, deleting agents, and viewing activity history to understand how each agent performs over time.

Follow the steps below to manage existing Agents in your project.

1. Go to Project Settings → AI Agents configuration to view your configured agents. Each agent overview card provides the following details:
   • Agent type - SEO expert, Translation, or Content summarizer
   • Workflow and workflow step where it runs
   • Number of seats
   • Agent configuration
   • Agent status
2. In the agent card, click the pencil icon to edit the agent. You can update:
   • The agent name and description.
   • The number of seats to adjust concurrent tasks.
   • The workflow or workflow step.
   • Any tone, locale, or custom instruction fields, based on agent type.
3. In the agent card, click Recent activity to see logs of recent agent runs. Use this data to monitor agent performance.
4. In the agent card, click the three dots context menu, and then click Delete to remove an agent. Deleting an agent removes it from all workflows, and queued or pending runs are automatically canceled.

#Best practices

Agents bring intelligent automation to your workflows. By combining context-aware AI with your content model, teams can effortlessly automate translation, summarization, and SEO. This reduces manual work, improves efficiency, and maintains editorial excellence. Follow these best practices to get the most consistent, high-quality results from your Agents.

• Set agent seats based on expected volume to balance performance and cost.
• Always validate AI-generated content for accuracy, tone, and keyword alignment before publishing.
• Position your agents strategically in workflows. Run these the agents during the Pre-publish or Optimization steps to ensure that operate on finalized content and produce the most accurate, high-quality results.
• For multilingual projects, pair the SEO expert with a Translation Agent to provide feedback on localized SEO metadata.