DeepResearch provides four primary output formats for its reports: plain text, Markdown, JSON, and HTML. These options cater to different developer workflows, ensuring compatibility with common tools and use cases. Each format is designed to deliver structured information while maintaining simplicity, allowing users to integrate results directly into their projects without unnecessary complexity.
Plain text is the most basic format, offering raw, unformatted content suitable for quick reviews or command-line processing. For example, developers can pipe plain text reports into scripts for automated analysis or log aggregation. Markdown outputs include lightweight formatting like headers, lists, and code blocks, making them ideal for documentation platforms (e.g., GitHub Wikis) or tools like Jupyter Notebooks. JSON provides machine-readable structured data, which is useful for APIs or applications that require programmatic access to report metadata, results, or statistical summaries. HTML reports support styled visualizations and tables, enabling direct embedding into web applications or dashboards. For instance, a developer might generate an HTML report to display interactive charts in a React app.
The format selection is typically configurable via API parameters or command-line flags. For example, an API request to DeepResearch might include "format": "json" in its payload to retrieve data for a Python script that parses results using the json module. Similarly, a CLI command could use --format markdown to generate documentation-ready output. This flexibility ensures developers can choose the right format for their pipeline—whether they need human-readable summaries, automated data extraction, or web-ready visuals—without additional conversion steps.