Coding & Development
RAG.md ยท 26 documents
plan
Plans a five-milestone RAG implementation for an Android gallery app using EmbeddingGemma, LiteRT, and DJL for on-device search.
๐ Development Notes
Documents an AI agent server's architecture, bugs, routing flow, and production gaps for developers building similar systems.
๐ Retrieve โ Implementation Plan
Defines an 11-phase implementation plan for a full-stack RAG application with document ingestion, vector search, and LLM answer generation.
Socrates Development Plan
Lays out a four-phase build plan for an AI tutor with text, voice, diagrams, and vision features.
Building a RAG Pipeline from Scratch
Teaches how to build a modular RAG pipeline from scratch with chunking, embeddings, vector storage, retrieval, and generation.
Multilingual Agentic RAG (venv, Qdrant, NiceGUI, Gemini-001)
Defines a multilingual agentic RAG pipeline using Gemini embeddings, Tesseract OCR, Qdrant, and NiceGUI with CPU-first defaults.
Partie 2 : Pipeline RAG
Guides a student through forking, cloning, Dockerizing, and running a RAG chatbot with Streamlit, ChromaDB, and Google Gemini.
The Simpsons Episode Database - Technical Plan
Plans a full-stack app that transcribes Simpsons episodes, creates searchable embeddings, and provides a web interface for semantic search and clip creation.
DinAI - RAG AI Chatbot Project Plan
Plans a four-phase RAG chatbot project, starting with a sequential web scraper for savollar.islom.uz and PostgreSQL/pgvector storage.
RAG (Retrieval Augmented Generation) Implementation - Technical Outline
Outlines RAG pipeline components, query translation techniques, indexing strategies, and adaptive systems for combining custom data with LLMs.
Implementation Plan: codeindex
We're building a local semantic index for codebases that augments Claude Code's Glob/Grep/Read tools with embedding-based search. The design is informed by OpenViking (AST skeletons, bottom-up directory summaries, score propagation), Augment Context Services (single retrieval tool for agents), sigma-ralph-grindset (`claude --print` as LLM backend), and db-harness (Bun-native dual-database skill pattern).
Loop #1 โ RAG Pipeline
Documents a two-loop system: a YouTube transcript RAG pipeline and a Pydantic AI agent with a FastAPI streaming API for querying the ingested content.
packrat (development version)
Documents changes across 15 versions of the Packrat R package for reproducible dependency management.
On Mac/Linux
Provides a complete Python script demonstrating seven LLM-powered workflows including email generation, venue booking, product ideation, FAQ chatbot, image prompts, and blog post creation.
RAG-Lighter Wiki
Documents a modular Python RAG framework with support for multiple LLMs, vector stores, embeddings, and evaluation.
How to set up QMD for Claude
Walks through installing qmd, indexing projects, adding it as an MCP server, and configuring Claude to use it before reading files.
Development Diary
Tracks development of a document similarity server with FastAPI, Typer CLI, and sentence embeddings, listing completed phases and usage examples.
Development Notes - AI Agent Server
Documents the development process, bugs, agent routing logic, and technical highlights of an AI agent server built with Express.js.
COSMIC Command Line Interface
Documents every command, option, and workflow for the COSMIC CLI tool used to chunk documents with configurable pipeline stages.
Technical Documentation - NanoScope
Documents the architecture, RAG pipeline, fine-tuning process, and performance metrics of a Chrome extension for AI-powered bookmark management and text manipulation.
Markdown VDB โ Technology Decisions
Justifies each Rust crate choice for a markdown vector database, linking requirements to specific libraries.
extend_retrieval
Explains how to customise Terrier's retrieval pipeline, weighting models, matching strategies, and query expansion, plus code snippets for index access.
Quickly build Generative AI applications with Amazon Bedrock
Provides runnable code samples for image generation, text tasks, chatbots, and RAG using Amazon Bedrock foundation models.
softrag
Provides a local-first RAG library using SQLite with sqlite-vec for document, embedding, and cache storage in a single.db file.