Search a folder of PDFs — and jump straight to the evidence
You know it's in there somewhere: a clause, a number, a sentence you half-remember, buried across dozens of PDFs. Opening them one by one and hammering Ctrl+F doesn't scale. Mempipe indexes the PDFs and Markdown notes in a Space so one search covers them all, and every result opens the exact passage in the source.
Exact when you need proof, ranked when you need recall
Different searches want different guarantees. Sometimes you need every occurrence of “Section 7.2” or a part number, matched exactly. Sometimes you want the passages that match “pricing risk” most closely, best first. Mempipe gives you both over the same corpus:
- →Grep-grade exact search across your PDFs and notes, with the precision of a developer tool.
- →Ranked full-text search when you know some of the words but not the exact phrase, best match first.
- →Path and folder scoping. Point Grep at one file, a glob pattern such as
contracts/**, or the whole Space. Ranked search can be limited to one file or one file type.
Results that end at the evidence
A search result that dumps you at the top of a 60-page PDF hasn’t finished the job. Because Mempipe indexes source-mapped documents, every result carries an anchor to its exact location — open it and you land on the passage itself, page and highlight included. The distance from “found a match” to “reading the evidence in context” is one click.
That same anchor is what powers citations everywhere else in Mempipe: citations in grounded Q&A answers, notes, and agent responses point at the same source locations search finds. One index, one provenance model, every surface.
Search your agents can use too
Every Space exposes its search over MCP — the same exact and full-text queries, the same source anchors. An agent hunting for evidence in your corpus works with the identical index a human uses in the workspace, so you can verify what it finds, and a valid citation from the agent opens the passage.
Learn how
Search inside your PDFs and Markdown files →
Open Search in a Space, choose BM25, and type a few words to list the matching passages from every PDF and Markdown file in the Space, best match first.
Find an exact phrase in your documents →
Use Grep in the Search panel to list every passage in your PDFs and Markdown files that contains an exact word or phrase, with options for letter case and patterns.
Open a citation and check its source →
Click a citation in an agent answer, a note, or a wiki page to open the source file at the cited passage, with a PDF's original page beside it and the passage outlined.
Frequently asked questions
- What kinds of search does Mempipe support?
- Two complementary kinds: exact, grep-grade search for when you know the phrase, identifier, or number you're hunting; and ranked full-text search for when you know some of the words but not the exact phrase. Both run over the same per-Space index.
- Which file types are searchable?
- PDF and Markdown files. Mempipe extracts each PDF into a structured, source-mapped document and indexes it alongside your Markdown notes, so one search covers your sources and your notes together. CSV and TSV files aren't searched; you query them with SQL instead.
- Do search results show where the match is in the PDF?
- Yes. Every indexed passage carries a source anchor back to its location in the original document. Opening a result lands you on the exact passage, with the page and highlight in the PDF.
- Can AI agents use the same search?
- Yes. Spaces expose search, grep, glob, and read tools over MCP, so agents query the same index humans use — and get results with the same source anchors, ready to cite.
Explore more
Build knowledge that comes with receipts.
Bring your sources. Find the evidence. Leave behind something worth keeping.
Public beta. Start on Free and upgrade to Starter for more capacity. Bring your own API key for the in-app agent. Reading, searching, and editing files needs no key.