One tool takes a block of text and returns semantically close chunks; the other takes a plain question and runs an agentic search across Slack, Gmail, Dropbox and Drive, and remember_this writes back into the same store. It points at the public Common Knowledge repository the project publishes, and the author calls the build in development — so treat the corpus as theirs rather than your team's.
An MCP server over a Rememberizer personal or team knowledge repository. The repository pulls from connected sources — Slack discussions, Gmail, Dropbox documents, Google Drive documents and uploaded files — and this server searches across them.
- retrieve_semantically_similar_internal_knowledge — send a block of text, get back the closest-matching chunks, optionally bounded to a date range
- smart_search_internal_knowledge — ask in plain language and let an agentic search assemble the result, with room to pass extra conversation context
- list_internal_knowledge_systems — which sources are actually connected to the repository
- list_personal_team_knowledge_documents — a paginated list of every document the knowledge system holds
- rememberizer_account_information — the account holder's name and email address
- remember_this — saves a named piece of text into the knowledge system so the two search tools can find it later
- Resources cover two kinds of material: documents and Slack discussions
Launched as a local process from the published package, or installed through the SkyDeck AI Helper app by searching for Rememberizer. This build points at the project's public Common Knowledge, which its README describes as being there for testing, and says the functionality is still in development and may change.