Using Datastreams as Agent Inputs

Once you've created and configured a Data Stream, the next step is to connect it to the Agents that will use it. This enables your content-generating Agents — such as blog writers, newsletter creators, or listicle generators — to draw from live, filtered, and relevant source material during every run.

Instead of relying on static prompts or generic search, Agents powered by Data Streams start with real-world context, increasing both accuracy and editorial value.

Where You Can Use Data Streams

Data Streams can be attached to any Agent that accepts a source input. Common examples include:

• Blog Post Generator
• Listicle Generator
• Newsletter Writer
• Social Post Composer
• Headline Generator
• SEO Optimizer
• Custom Agents with an {{Input_Source}} field

Whether the Agent is writing long-form content, composing summaries, or generating metadata, a linked Data Stream ensures the output is grounded in relevant, up-to-date material.

How to Assign a Data Stream to an Agent

Step-by-Step:

1. Open the Agent Configuration Panel
   Navigate to any Agent template or instance and open its settings.
2. Locate the Input_Source Field
   This is the field where content is pulled from. It may be labeled as “Source Content,” “Primary Input,” or “Research Base” depending on the Agent.
3. Select a Data Stream
   From the dropdown or Data Stream browser, select the Stream you want the Agent to use.
4. Save and Run the Agent
   When you trigger the Agent, it will now pull its context directly from the latest articles or topic clusters inside the Data Stream.

How the Integration Works

When a Data Stream is used as an input:

• The Agent receives a live article or cluster summary that matches the Stream's filters
• The selected content is automatically injected into the {{Input_Source}} placeholder at runtime
• The Agent prompt runs as usual, now enhanced by real-world data
• If configured, the Agent may also use quote tagging, citation logic, or metadata mapping built into the Stream

This allows each output to reflect not just brand tone — but editorial reality.

Input Behavior Based on Stream Type

Agents that require deep research, citations, or quote verification should use Article-type Streams.
Agents that focus on tone, style, or top-level concept generation may prefer Inspiration-type Streams.

Optional: Quote Logic and Metadata Handling

Advanced Agents may include special options tied to Data Stream behavior:

• Quote inclusion toggle – whether to auto-insert quotes from source articles
• Metadata filters – allow Agents to reference specific tags, authors, or source types
• Length and freshness filters – Agents can be instructed to prioritize short, recent, or in-depth articles

These settings are defined in the Agent prompt or its configuration logic.

Things to Watch

• Stream must be published – Draft or inactive Streams will not populate Agent inputs
• Ensure your filters are effective – An empty Stream may result in no content or fallback behavior
• Use the Preview tool – Always test the Stream first to verify the type and quality of content it delivers
• Multiple Agents can share one Stream – A single well-tuned Stream can be reused across many content formats

Summary

Using Data Streams as Agent inputs transforms your content creation process from static prompting to live editorial execution. It brings in fresh ideas, trusted facts, and structured context — all automatically aligned with your brand’s direction.