Mempipe terms explained
This glossary explains the words you meet in the Mempipe app and in this help, in plain language. Each entry links to the page that shows you how to use it. Words in bold inside an entry are labels the app shows.
Your account
Section titled “Your account”Organization
Section titled “Organization”An organization holds your Spaces and your plan. You create your first one when you sign up, and you can own up to three. The app has no way to add a member besides the person who created it. See Create and switch organizations and Spaces.
A Space holds a set of related files, such as the sources and notes for one project, and everything Mempipe makes from them. Search, the in-app agent, and each AI app you connect work in one Space at a time. See What is a Space?
A slug is the short name of an organization or a Space in its web address, such as acme-research and research in app.mempipe.ai/@acme-research/research. You choose it when you create the organization or Space, and you type it again to confirm a deletion. It can’t be changed later.
An owner can change an organization’s plan and delete the organization or its Spaces. The person who creates an organization becomes its owner. See Delete a Space, an organization, or your account.
A plan sets an organization’s limits on storage, files and folders, Spaces, and the size of each file. Mempipe has two plans today, Free and Starter, and each organization has its own. See Plans, usage limits, and upgrading.
Home lists your organizations. Open it from the Mempipe logo menu at the top left of the app. See Switch organizations.
Files and processing
Section titled “Files and processing”Files is the panel that lists a Space’s files and folders. Open it from the narrow rail at the left edge of a Space. See Upload files to a Space.
Source
Section titled “Source”A source is a file you collect and learn from, such as a PDF report or a paper. In an LLM wiki, sources go in the raw folder and stay unchanged. See What is an LLM wiki?
Processing
Section titled “Processing”Processing is what Mempipe does after you upload or save a file. It extracts the text of a PDF or Markdown file so you can read, search, and cite it, and it reads a CSV or TSV file into a table. The app calls this deriving: a file shows Deriving… while it runs, then Ready, or Derive failed or Search failed if a step fails. See What happens after you upload a file.
Extractor
Section titled “Extractor”An extractor is the tool that turns a PDF into Markdown text. Docling is the default. LlamaParse is the other choice, for a PDF whose text comes out wrong. See Choose an extractor.
A view is one way of showing a file. Choose it in the View switcher above the file:
- original shows the file as you uploaded or saved it.
- markdown shows a PDF’s extracted text.
- reading shows a Markdown file formatted, with working links.
- tables, schema, and query show a CSV or TSV file as a table, describe its columns, and run SQL over it.
- visualization draws a chart file.
- search-units lists the passages of a PDF or Markdown file that search can find.
Passage
Section titled “Passage”A passage is one paragraph, heading, list item, table row, or code block of a file’s text. Search results and citations point to passages.
Page chip
Section titled “Page chip”A page chip, such as p.12, marks where a passage sits in a PDF. Choose it to open the original page with the passage outlined. The number counts pages from the first page of the file, so it can differ from the number printed on the page. See Read a PDF as Markdown and check the original page.
stale beside a search result or on a view means the file has changed since Mempipe last processed it, so what you see may be out of date. stale — re-deriving means the new version is being processed now. refresh failed means processing the new version failed. See If results are out of date.
Wiki link
Section titled “Wiki link”A wiki link is a file’s name in double square brackets, such as [[overview]], inside a Markdown file. In the reading view, it opens that file. With ! in front, such as ![[overview]], it shows the file inside the note. See Link to other notes.
Space Activity
Section titled “Space Activity”Space Activity lists what happened in a Space: uploads, edits, renames, moves, deletions, and processing steps, with the reason a step failed. Open it from Pipeline. See See what happened in Space Activity.
Pipeline
Section titled “Pipeline”Pipeline is the icon in the left rail that opens Space Activity, Search Index, and MCP Endpoint. See Find your way around a Space.
Inspector
Section titled “Inspector”Inspector is the tab in the right panel that shows the open file’s details, such as its path, type, size, and version. See Find your way around a Space.
Search and citations
Section titled “Search and citations”BM25 is the Search panel’s ranked mode. It lists the passages that contain the words you type, best match first. It matches words, not meanings. See Search inside your PDFs and Markdown files.
Grep is the Search panel’s exact-text mode. It lists every passage that contains a word, phrase, or pattern, sorted by file. See Find an exact phrase in your documents.
Glob is the Search panel’s file-name mode. It lists every file whose path matches a pattern, such as raw/*.pdf. See Find a file by name.
A regex, or regular expression, is a short code that describes text to match, such as risk (tolerance|appetite). Grep reads its pattern as a regex while Regex is on. See Search, grep, and glob syntax.
Search Index
Section titled “Search Index”Search Index is the page that lists every passage search can find in a Space. Use it to check whether a file has reached search. Open it from Pipeline. See everything search can find shows how to use it.
Citation
Section titled “Citation”A citation is a link to the passage that an answer, note, or wiki page relies on. Choose it to open the source file at that passage, with the passage highlighted. A citation shows where the words came from. It doesn’t prove the claim beside it is right. See Open a citation and check its source.
Agents and AI apps
Section titled “Agents and AI apps”In-app agent
Section titled “In-app agent”The in-app agent is the AI assistant in a Space’s Agent tab. It searches and reads the Space’s files, answers with citations, and saves notes, queries, and charts when you ask. It uses an API key you bring. See Ask the agent about your files.
Model provider and API key
Section titled “Model provider and API key”A model provider is the company whose AI model the in-app agent uses: OpenAI, Anthropic, Vercel AI Gateway, or OpenRouter. An API key is the secret code from that provider that lets the agent use your account there. The provider bills you for what the agent uses. See Set up the in-app agent with your own API key.
AGENTS.md
Section titled “AGENTS.md”AGENTS.md is a Markdown file at the top level of a Space that holds standing instructions for the in-app agent. The agent reads it before every message you send in that Space. See Give the agent standing instructions with AGENTS.md.
MCP, the Model Context Protocol, is the open standard AI apps use to connect to outside tools and data. Apps such as Claude, Claude Code, ChatGPT Desktop, Codex, Cursor, and VS Code use it to search and read a Mempipe Space. See Connect an AI assistant to Mempipe with MCP.
MCP Endpoint
Section titled “MCP Endpoint”MCP Endpoint is the page with a Space’s MCP address and the setup steps for each AI app. Open it from Pipeline. See Connect an AI assistant to Mempipe with MCP.
Write access
Section titled “Write access”Write access lets a connected AI app create and replace Markdown, SQL, and chart files in a Space. An app gets it only if you tick the box for it when you approve the connection. On the all-Spaces connection, it also lets the app add PDF and CSV files from ChatGPT and delete single files. No connection can move files or delete a folder. See What a connected AI can read and change.
Tables, SQL, and charts
Section titled “Tables, SQL, and charts”mp_table
Section titled “mp_table”mp_table('path') names a CSV or TSV file inside a SQL query, such as SELECT * FROM mp_table('data/sales.csv'). See Query CSV files with SQL.
Saved query
Section titled “Saved query”A saved query is a .sql file. It runs again each time you open it, and notes and charts can show its result. See Save a query as a .sql file.
Chart file
Section titled “Chart file”A chart file is a .vl.json file that describes a chart in Vega-Lite, an open charting format. Its data comes from a CSV file or a saved query in the same Space. See Make a chart from your data.
LLM wiki
Section titled “LLM wiki”LLM wiki
Section titled “LLM wiki”An LLM wiki is a set of linked Markdown pages that an AI agent writes and keeps up to date from your sources. See What is an LLM wiki?
raw and wiki folders
Section titled “raw and wiki folders”In an LLM wiki, the raw folder holds your sources, and the wiki folder holds the pages the agent writes. wiki/index.md lists the pages, and wiki/log.md gets a short entry each time a source is added or the wiki is tidied. See Build an LLM wiki from your PDFs.