Skip to content

About

Personal knowledge system with RAG Agent, linked to the Obsidian vault, rendered to the Static Web

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

DevBook product overview

A structured collection of software engineering notes covering 11 topic areas, designed for learning and interview preparation at a Senior .NET / AI engineering level.

Read online Latest version CI status Report an issue Contributing License

Get started

Browse the published topic pages. The collection covers 11 areas: programming, computer science, data persistence, networks, architecture, development practices, security, SDLC, DevOps, cloud, and AI & ML. Notes support learning, recall, and interview preparation at a Senior .NET / AI engineering level.

The site publishes the eligible subset of the Obsidian vault. Reading does not require Obsidian, a database, or backend credentials.

Local setup

Use the Obsidian vault

Clone the repository and open Vault/ as a vault in Obsidian:

git clone https://github.com/grafanaKibana/devbook.zip.git
cd devbook.zip

Start with Vault/Home/index.md. Notes and interview cards are authored within their owning topics.

For local authoring with the checked export workflow, install Git, Bash, .NET SDK 10.0.100, Node 22/npm, and Obsidian desktop. Allow the bundled community plugins, enable/register Settings → General → Command line interface, keep this vault open, then run:

bash .scripts/setup.sh

This installs locked Web dependencies, enables the checkout's Git hooks, and configures Quartz Syncer's local destination with Preserve enabled. Setup does not export, stage, commit, or deploy. Git identity and GitHub authentication remain manual, and Datacore must finish indexing before committing notes.

The workflow is tested on macOS; Linux is unverified, and native Windows and mobile export are unsupported. The local development guide covers the bundled Syncer build, prerequisites, and recovery.

Run the RAG playground

Platform/DevBook/ is a personal .NET 10 learning proof of concept for retrieval and agent mechanics. It exposes ingestion, vector search, and an ask endpoint over the notes, with separate retrieval evaluations.

Before running it, follow the backend prerequisites: the host requires MongoDB Atlas search support and OpenAI configuration. Keep credentials in user-secrets or environment variables. Then, from the repository root:

dotnet run --project Platform/DevBook/DevBook.API/DevBook.API.csproj

Ingest publishable notes before trying retrieval. This playground makes the mechanics visible; it is not a production assistant. Evaluation measures retrieval of evidence, rather than grading generated answers.

Repository structure

Path Role
Vault/ Authored Obsidian notes, templates, and attachments
Web/ Quartz v5 site and generated publication output
Platform/ RAG backend, tests, and retrieval evaluations
.scripts/ and .githooks/ Vault maintenance and checked export workflow

A vault commit waits for Syncer export; a pushed change then follows the website's build/deployment workflow. Edit Vault/, not generated Web/content/.

Guides

About

Personal knowledge system with RAG Agent, linked to the Obsidian vault, rendered to the Static Web

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Contributors

Languages