add-parallel

You can integrate web search into NanoDex by configuring the container and agent runner scripts with a Parallel AI API key, enabling real-time information retrieval and deep research task execution directly within the assistant workflow.

Deep research works by utilizing a dedicated task API to conduct in-depth analysis on complex topics. It operates non-blocking through NanoDex's scheduler, allowing the main workflow to continue while fetching comprehensive results.

Yes, a valid Parallel AI API key is required. You must update the container runner and agent runner scripts configuration to authenticate requests and enable both quick web search and deep research capabilities.

Yes, non-blocking web searches are supported through NanoDex's scheduler. This mechanism allows the assistant to fetch real-time data or conduct deep research asynchronously, avoiding interruptions to the primary workflow.

Quick web search performs fast, factual lookups for current information, while deep research conducts in-depth analysis on complex topics using a dedicated task API. Both operate non-blocking via NanoDex's scheduler.

The primary limitation is dependency on the scheduler's capacity and external API rate limits. Complex deep research tasks may take longer to return results, although they run non-blocking to prevent main workflow interruption.