Skip to main content
Recall is the payoff. Everything you capture and file exists so that later, when you or an agent needs it, it comes back. Cortex finds things by meaning, not just by matching words.

​
Search understands what you mean

Ask a brain “what did we decide about pricing?” and it does not just hunt for the word “pricing”. It understands the intent and returns the pages that are actually about that decision, even if they phrase it differently. This is what makes recall feel like asking a knowledgeable colleague rather than running a keyword search. You can search a brain yourself from the app. Your connected agents search through their connection, across all the brains you can access at once, and every result says which brain it came from, so wide recall never blurs where anything lives.

​
What makes recall good

The quality of what comes back depends on how well things were filed, which is the whole reason structure matters:
  • Citations mean answers arrive with their sources, so you can trust them and check them.
  • Links mean that finding one page often surfaces the related ones around it.
  • Clean pages mean the search is reading tidy write-ups, not raw transcripts, so it finds the substance.
A brain that has been filed well recalls well. That is why Cortex puts effort into filing: it is really an investment in recall.

​
Recall through your agent

The most powerful way to use recall is through an agent that is connected to your brain. Instead of searching yourself and reading results, you ask the agent a question and it does the searching, reads the pages, and answers you, with citations. Over time, as your brain fills up, the agent gets more useful, because it has more to draw on.

​
Try it

  1. Search a brain with a plain-language question rather than keywords.
  2. Open a result and follow its citation back to the source.
  3. Ask a connected agent the same question and compare how it answers.