Local knowledge automation
Synapse is a local notes and RAG lab: Markdown notes are structured, made visible in a wiki, and queried through a terminal client.
Synapse connects technical Markdown notes to a local search and question-answer pipeline. Notes remain readable as simple text files, are also published to a wiki, and are indexed for semantic search. The Ask TUI lets you query your own notes and shows the source behind an answer.
A learning note, lab document, or runbook kept as a simple Markdown file.
Synapse formats and processes the note for wiki display and indexing.
Wiki.js keeps the content readable while Qdrant makes it semantically searchable.
The terminal client queries local knowledge and shows answers with source context.
Ask / terminal workflow
The GIF shows Synapse Ask in a terminal: an OSPF question is answered from indexed notes, with the answer pointing to its source. It demonstrates the core idea: local notes are not only stored, but made practically queryable.
Setup steps, ports, service relationships, error messages, and technical decisions can be queried from Markdown notes instead of searched manually through old files.
Technical lab knowledge stays findable without a heavy documentation system.
Recurring service checks, troubleshooting steps, and local setup procedures can be maintained as simple notes and queried directly in the terminal.
Repeatable admin workflows stay local, simple, and quickly accessible.
Synapse keeps project notes simple as Markdown while making them more useful through a wiki and semantic search.
Project knowledge remains maintainable as Markdown and becomes easier to search.
Answers should visibly rely on the indexed notes. Source context makes it clear where an answer came from.
The value is traceability, not free-form chat without an auditable basis.
Synapse was mainly an exercise in making documentation practically useful. Notes only have value when they can be found, understood, and used later.
A good RAG project becomes stronger through a traceable flow: source in, processing visible, content discoverable, answer returned with a reference.
This is a local lab system, not a production-ready knowledge platform. That keeps the focus clear: local notes, a traceable workflow, and simple terminal use.
The local note, indexing, and Ask workflow is in place and documented as a traceable flow.
The project is a local lab, not a production environment. Its features, model behaviour, and complete stack are shaped around that local use case.
