Content research powered by semantic search

Discover and find what you need with natural language on thousands of content pages

How does it work

Build your personal AI library

Gather every content source you have into an AI library – our semantic engine can digest all types of sources – books, audio, video, you name it.

Any source you add to your library gets fragmented into smaller pieces that are linked to each other based on how they appear.

Use natural language

You can search and find relevant text fragments by doing a semantic search. Use natural language to search for what you are looking for. Only those fragments with similar meanings will appear.

Additionally, you will see a similarity score and where does it appear within the source. For example, for webs the fragment source will be the URL of the page.

Follow up on content

Use the controls to expand or collapse text fragments to see what’s right before or after as it appears in the cited source.

Extract insights

Use your search results as informational context to run AI prompts to summarize or apply any action you want over them. This works especially well on language models with large context windows, such as [Claude AI](

What sources can be ingested


Any websites is eligible for ingestion. Our engine will crawl every page within a particular set of constraints that you specify, such as domain, starting URL or “must contain a particular term”.


Any audio or video based source can also be ingested, whether it’s a file or hosted on other sites like YouTube. We will use the captions or transcribe the audio to index its content as text.

Text and books

You can also ingest digital book formats like epub or other common formats such as PDF. Structured formats such as the ones used in ebooks will work best as they already come highly formatted.

Request new ingestors

We keep enhancing our library of ingestors so you can get in your AI library what you need. The more content you have in it, the more powerful your searches are.

Using your library content

Use search to enhance prompts

You can use the results of a library search as informational context in AI prompts. This is a powerful way to grab any generic language model such as ChatGPT, and make it become an expert o

Imitating particular writing styles

You can use the content in your library to extract style traits that define how those fragments were written. Once you analyze and gather those traits into an AI Persona, you can reuse it to generate new AI content using that same style.

Use search results to create an AI Persona →