Search site

Type a word or phrase to search public pages.

    Files as graph values

    Store any type of file as a value in the graph, connected to the facts around it.

    Use the graph model you know

    Files are graph values alongside strings, dates, and numbers. Legra's extended value model works with RDF and labelled property graphs (LPG).

    Cypher and GQL

    A vertex (node) can have a property whose value is an entire file. For a meeting recording:

    VertexPropertyValue
    A meetingRecordingAn audio file

    If the recording has its own author, date, or other metadata, give it a separate vertex with a content property. Connect that vertex to the meeting with a has-recording edge.

    RDF and SPARQL

    An RDF triple expresses the same example as a subject, predicate, and object. Legra's extended RDF literals let the object contain the audio file:

    SubjectPredicateObject
    A meetingHas recordingAn audio file literal

    Both views use the same file content and media type.

    Any media type

    Keep Word documents, PDFs, spreadsheets, email, messages, images, audio, video, and other file types in the graph. Connect them to customers, authors, projects, or decisions through ordinary graph relationships.

    Large files stream into and out of storage without requiring the complete value in memory. File content follows the workspace's access and encryption rules and version history.

    Save space through deduplication

    Legra divides large content into reusable chunks, taking file structure into account. Identical chunks within a workspace are stored once and shared across files and versions. This deduplication lets a revised document reuse the parts that did not change.

    Add folders with Drive

    File values are a platform capability available to every app. Drive adds directories, filenames, file versions, and connections to external file stores such as SharePoint.

    Whole websites fit as collections of HTML, CSS, JavaScript, images, and other file values. Drive keeps their directory structure alongside their content in the graph.

    Developer reference

    Technical details

    Value and storage model

    File bytes are content-bearing literals in Legra's extended RDF value model or property-graph properties. An IANA media type identifies representation; metadata triples identify filename, author and source. This binary literal extension is distinct from standard RDF datatypes.

    The storage path streams chunks into encrypted, content-addressed workspace storage. Retrieval reconstructs the exact byte sequence; physical chunking does not alter logical value identity. Chunk reuse stays inside the workspace encryption boundary.

    OCR text, embeddings and thumbnails are separately derived artifacts. Each needs its own producing task and provenance; storing application/pdf does not execute an extraction pipeline.