Memory

Browser memory stores pages you've visited and notes you've written. Store now, recall later.

Remember a page

omniscout remember https://example.com/blog/post

remember visits the URL, extracts its content, and stores it. Search it later with memory:

omniscout memory "neural networks in production"

memory ranks stored pages by similarity to your query, so paraphrases match — you don't need the exact title or URL.

Search your own memory first

The search command can target memory instead of the web:

omniscout search "query" --source memory

Useful in agents: check what you've already seen (--source memory) before spending time on a fresh web search.

Manage the corpus

memory is a command group, not just a query:

omniscout memory "query"     # search remembered pages
omniscout memory list        # list records
omniscout memory show <id>   # show one record
omniscout memory note "idea: compare options"   # standalone note
omniscout memory delete <id> # delete records
omniscout memory stats        # corpus statistics
omniscout memory clear        # wipe the entire corpus

Attach a note to the active page (or write a standalone one) with memory note — notes are indexed too, so decisions and context resurface in later searches.

Related commands

Related commands:

omniscout index ...          # manage the search index
omniscout crawl ...          # run and inspect crawl jobs
omniscout extract-jobs ...   # inspect background extract jobs

Memory, saved pages, and caches all live in the data dir — see Configuration. They never leave your machine; the only outbound traffic is searches you run and URLs you fetch.