Prototype

Deep research

A multi-source research engine that searches, scrapes, transcribes, and synthesises content into structured digests.

Enter a topic and it hits four sources in parallel: Perplexity AI for editorial deep dives with citations, ScrapingBee for Google Search and news results, YouTube for videos with auto-transcription, and Twitter for social signals. Results land in collapsible sections with an insights sidebar tracking volume metrics. Select any text to save it to a personal knowledge base. Supports time-range filtering from breaking news to all-time, and a detailed brief mode that auto-summarises complex queries for cross-source consistency. Digests are saved as timestamped JSON.