Find anything,
by keyword or by meaning.
Hybrid search combines exact keyword matching with meaning-based search across every file in your workspace. You can also query the structured metadata extracted from your documents.
Either party may end this agreement with thirty days written notice.
The termination notice period shall be sixty days.
One returns exact matches. The other understands intent.
Keyword search and semantic search answer different questions. Fastio runs both at once and ranks the results together, so you never have to pick the right tool before you start typing.
Keyword search
Full-text matching that finds the exact word or phrase you typed. When you know the precise term, a name, an invoice number, a clause title, keyword search returns it directly.
- Exact full-text matching across documents, spreadsheets, and notes
- Find a contract number, a clause, or a person's name in seconds
- Filenames and file contents are both indexed
Semantic search
Meaning-based search that understands intent. It matches files that express the same idea even when they use different words, so you find the right passage without guessing the exact phrasing.
- Search by meaning, not just the exact words you typed
- A query for "termination terms" surfaces "end of agreement" clauses
- Useful when you remember the idea but not the wording
Hybrid runs both and merges the results
Every query checks exact keywords and meaning at the same time, then ranks the strongest matches first. You get the precision of keyword search and the recall of semantic search in a single list.
Query the fields inside your files.
Metadata Views extract structured fields from your documents, the amount on an invoice, the jurisdiction in a contract, the renewal date in an agreement. Once those fields exist, you can search by their values, not just the text on the page. Ask for every invoice over a threshold or every contract governed by a specific state and get back the exact files that match.
- Filter on extracted fields: query the structured values, not just the raw document text
- Combine with keyword and semantic: narrow a meaning-based search down to files that also match a field
- Treats your files like a database: sort and filter a folder of documents by the data they contain
amount > 10000 Return every invoice over a threshold without opening one file
jurisdiction = California Filter contracts by an extracted governing-law field
renewal_date before 2026-09-01 Pull agreements that come up for renewal this quarter
status = unsigned List documents still waiting on a signature
Search finds the files. Ripley answers the question.
Search and Ripley AI chat solve two halves of the same problem. Use search when you need to locate and open the exact file. Use Ripley when you want an answer written for you, with citations back to the sources it used.
Search returns the matches
You get a ranked list of files and the highlighted passages inside them. You open the result and read the source yourself. Ideal when you know the document exists and just need to put your hands on it.
Ripley answers and cites
Ask Ripley a question and it reads the relevant files, writes a plain-language answer, and links each point back to the document it came from. Ideal when you want a conclusion, not a reading list.
The same index powers both. Agents reach search and Ripley through a consolidated MCP toolset, so an automation can locate files and ask questions about them in one workflow.
Stop scrolling. Start finding.
Keyword precision, semantic recall, and metadata filtering across every file in your workspace. Search locates the source, and Ripley answers from it.