OpenLoci

OpenLoci is a filesystem-native memory palace generator designed to facilitate distributed cognition and structured project environments. Drawing on the Method of Loci as both metaphor and architecture, it organizes information by mapping conceptual spaces to the filesystem: each room corresponds to a directory, and each artifact is represented by a markdown file. The tool emphasizes collaborative use, with every agent session leaving a trace, turning interactions into a navigable walk through a virtual space.

OpenLoci supports both finite and infinite project types. For finite projects—such as investigations or migrations—the structure provides clarity and a clear endpoint. For ongoing, infinite practices like research or writing, the palace grows organically as it is used. The platform introduces the concept of "skins," which alter the cultural frame of the palace without changing its underlying structure; examples include themes inspired by Clue, The X-Files, Silicon Valley, The Muppets, and The Amazing Digital Circus. Users can also author their own skins.

The generated project environments include specialized directories for configuration, agent profiles, guidelines, journaling, and various workstreams such as research, collaboration, infrastructure, and retrospectives. A built-in Journal feature allows for regular blogging about sessions, providing narrative cohesion to ongoing work. OpenLoci is built with minimal computing principles, relying on stable technologies like filesystem directories and plain text files, and is platform-agnostic. It is compatible with any Large Language Model that can access the filesystem, making it adaptable to various workflows.

From a technical perspective, OpenLoci leverages YAML frontmatter for metadata and routing, Cookiecutter for templating project structures, Hugo for rendering and navigation, Git for version control and history, and a command-line interface built with Typer and Rich for user interaction. The software enforces a constraint that significant outputs must be grounded in material receipts—such as notes, evidence, or summaries—ensuring that meaningful work is always recorded. 0 license and can be installed via pip.