Resource, examples & tutorials for multimodal AI, RAG and agents using vector search and LLMs
974
stars
906
commits
Jupyter Notebook
primary language
Apr 24, 2026
updated
README
VectorDB-recipes
Dive into building GenAI applications! This repository contains examples, applications, starter code, & tutorials to help you kickstart your GenAI projects.
- These are built using LanceDB, a free, open-source, serverless vectorDB that requires no setup.
- It integrates into Python data ecosystem so you can simply start using these in your existing data pipelines in pandas, arrow, pydantic etc.
- LanceDB has native Typescript SDK using which you can run vector search in serverless functions!
Join our community for support - Discord • Twitter
This repository is divided into 2 sections:
- Examples - Get right into the code with minimal introduction, aimed at getting you from an idea to PoC within minutes!
- Applications - Ready to use Python and web apps using applied LLMs, VectorDB and GenAI tools
The following examples are organized into different tables to make similar types of examples easily accessible.
Sections
- Build from Scratch - Step-by-step guides to create AI applications from scratch.
- Multimodal - Build apps that process and search across both text and images.
- RAG - Combine document retrieval with LLM-powered responses.
- Vector Search - Learn to efficiently find relevant documents using vector-based search.
- Chatbot - Create AI chatbots that fetch information and generate intelligent replies.
- Evaluation - Measure the quality and accuracy of AI-generated answers.
- AI Agents - Build LLM-driven applications where multiple agents collaborate and interact.
- Recommender Systems - Develop AI-powered recommendation systems for personalized suggestions.
- Concepts - Tutorials and explanations of key techniques used in AI applications.
🌟 New 🌟
Stay up to date with the latest projects, tools, and improvements added to the repository.
- V-JEPA Video Search -
Build from Scratch
Start with the basics! These examples guide you through creating AI applications from the ground up using LanceDB for efficient document retrieval and search.
| Build from Scratch | Interactive Notebook & Scripts |
|---|---|
| Build RAG from Scratch |   |
| Local RAG from Scratch with Llama3 |   |
| Multi-Head RAG from Scratch | |
| Fintech AI Agent from Scratch |  |
MultiModal
Search across different types of data (text, images, and more). Build powerful search applications that work with diverse inputs.
RAG
Generated Responses by retrieving relevant documents before answering. This section covers different approaches to implementing RAG in your projects.
Vector Search
Find relevant documents quickly! These projects show how to use vector-based search techniques to make AI-powered searches faster and smarter.
Chatbot
Create chatbots that understand user queries and fetch relevant responses using LanceDB’s vector search capabilities.
Evaluation
These projects provide tools to compare AI-generated responses against reference data and fine-tune accuracy.
| Evaluation | Interactive Notebook & Scripts | Blog |
|---|---|---|
| Monitoring and Tracing RAG using HoneyHive | | |
| Evaluating RAG with RAGAs |  | |
AI Agents
Build applications where multiple AI agents interact to complete tasks efficiently. These projects show how agents can collaborate, exchange data, and automate workflows.
Recommender Systems
Personalized AI recommendations! These projects help you build recommendation engines that suggest content based on user preferences.
| Recommender Systems | Interactive Notebook & Scripts | Blog |
|---|---|---|
| Movie Recommender | | |
| Product Recommender | | |
| Arxiv paper recommender | | |
| Music Recommender | | |
Concepts
Learn the core ideas behind AI applications—including text chunking, retrieval strategies, and optimization techniques—to improve your understanding of vector search and AI pipelines.
Projects & Applications
Ready-to-use AI applications built with LanceDB! Use these projects as-is, customize them, or integrate them into your own applications.
Node applications powered by LanceDB
| Project Name | Description | Screenshot |
|---|---|---|
| Writing assistant | Writing assistant app using lanchain.js with LanceDB, allows you to get real time relevant suggestions and facts based on you written text to help you with your writing. | |
| Sentence Auto-Complete | Sentance auto complete app using lanchain.js with LanceDB, allows you to get real time relevant auto complete suggestions and facts based on you written text to help you with your writing.You can also upload your data source in the form of a pdf file.You can switch between gpt models to get faster results. |  |
| Article Recommendation | Article Recommender: Explore vast data set of articles with Instant, Context-Aware Suggestions. Leveraging Advanced NLP, Vector Search, and Customizable Datasets, Our App Delivers Real-Time, Precise Article Recommendations. Perfect for Research, Content Curation, and Staying Informed. Unlock Smarter Insights with State-of-the-Art Technology in Content Retrieval and Discovery!". |  |
| AI Powered Job Search | Transform your job search experience with this AI-driven application. Powered by LangChain.js, LanceDB, and advanced semantic search, it provides real-time, highly accurate job listings tailored to your preferences. Featuring customizable datasets and advanced filtering options (e.g., skills, location, job type, and salary range), this app ensures you find the right opportunities quickly and effortlessly. Best suited for job seekers, recruiters, career platforms, custom job boards. |  |
| AI Powered Multimodal meme search | An advanced AI-powered meme search engine that allows users to find memes using both text and image queries. By leveraging LanceDB as a high-performance vector database and Roboflow's CLIP model for embedding generation, the platform delivers fast and accurate meme retrieval. |  |
| AI Powered Feedback search and analysis | An AI-powered employee feedback analysis platform designed to collect, store, analyze, and retrieve insightful employee feedback. This system leverages LanceDB for high-speed vector-based semantic search, React.js for an interactive UI, Node.js for backend processing, and LangChain.js with an Ambient Agent for intelligent analysis and actionable insights. |  |
| Hierarchical Multi Agent | The AI-Powered Law Assistant is a Hierarchical Multi-Agent System leveraging LangGraph, LangChain, and LanceDB for efficient legal query processing. It features a Supervisor Agent that delegates tasks to specialized agents for IPC and NDPS laws, each with sub-agents for case retrieval and legal summarization. Using LanceDB, it stores and retrieves vectorized legal documents, enabling fast, structured, and context-aware responses for legal professionals, researchers, and law students. |  |
| Project Name | Description | Screenshot |
|---|---|---|
| YOLOExplorer | Iterate on your YOLO / CV datasets using SQL, Vector semantic search, and more within seconds | |
| Website Chatbot (Deployable Vercel Template) | Create a chatbot from the sitemap of any website/docs of your choice. Built using vectorDB serverless native javascript package. |  |
| Advanced Chatbot with Parler TTS | This Chatbot app uses Lancedb Hybrid search, FTS & reranker method with Parlers TTS library. |  |
| Multi-Modal Search Engine | Create a Multi-modal search engine app, to search images using both images or text | |
| Evaluate RAG | A working Streamlit RAG App designed to demonstrate end to to end production grade evaluation using 50+ scores and metrics which include guards, software metrics, traditional metrics and LLM as judge metrics. It uses mixture of specialised deep learning models and LLM as Judge models to do the evaluations |  |
| Multi-Agent Collaboration Chatbot | Multi-Agent collaboration chatbot using langgraph for share-market use case using Lancedb & tools such as Polygon ,Tavily |  |
| Multimodal Myntra Fashion Search Engine | This app uses OpenAI's CLIP to make a search engine that can understand and deal with both written words and pictures. |  |
| Multilingual-RAG | Multilingual RAG with cohere embedding & support 100+ languages | |
| Music Recommender | Music Recommendation system using audio feature extraction and vector similarity search. By utilizing LanceDB, PANNs for audio tagging, and Librosa for audio feature extraction, the system finds and recommends tracks with similar audio characteristics based on a query song. |  |
| NoOCR | End-to-end solution for complex PDFs, powered by ColPali and LanceDB. |  |
🌟 New! 🌟 Applied GenAI and VectorDB course on Udacity Learn about GenAI and vectorDBs using LanceDB in the recently launched Udacity Course
Contributing Examples
If you're working on some cool applications that you'd like to add to this repo, please open a PR!
Contributors
(top 30 of 37)
Similar repos
Languages
Jupyter Notebook
92.6%
Python
3.7%
JavaScript
3.0%
Resource, examples & tutorials for multimodal AI, RAG and agents using vector search and LLMs
974
stars
906
commits
Jupyter Notebook
primary language
Apr 24, 2026
updated
README
VectorDB-recipes
Dive into building GenAI applications! This repository contains examples, applications, starter code, & tutorials to help you kickstart your GenAI projects.
- These are built using LanceDB, a free, open-source, serverless vectorDB that requires no setup.
- It integrates into Python data ecosystem so you can simply start using these in your existing data pipelines in pandas, arrow, pydantic etc.
- LanceDB has native Typescript SDK using which you can run vector search in serverless functions!
Join our community for support - Discord • Twitter
This repository is divided into 2 sections:
- Examples - Get right into the code with minimal introduction, aimed at getting you from an idea to PoC within minutes!
- Applications - Ready to use Python and web apps using applied LLMs, VectorDB and GenAI tools
The following examples are organized into different tables to make similar types of examples easily accessible.
Sections
- Build from Scratch - Step-by-step guides to create AI applications from scratch.
- Multimodal - Build apps that process and search across both text and images.
- RAG - Combine document retrieval with LLM-powered responses.
- Vector Search - Learn to efficiently find relevant documents using vector-based search.
- Chatbot - Create AI chatbots that fetch information and generate intelligent replies.
- Evaluation - Measure the quality and accuracy of AI-generated answers.
- AI Agents - Build LLM-driven applications where multiple agents collaborate and interact.
- Recommender Systems - Develop AI-powered recommendation systems for personalized suggestions.
- Concepts - Tutorials and explanations of key techniques used in AI applications.
🌟 New 🌟
Stay up to date with the latest projects, tools, and improvements added to the repository.
- V-JEPA Video Search -
Build from Scratch
Start with the basics! These examples guide you through creating AI applications from the ground up using LanceDB for efficient document retrieval and search.
| Build from Scratch | Interactive Notebook & Scripts |
|---|---|
| Build RAG from Scratch | |
| Local RAG from Scratch with Llama3 | |
| Multi-Head RAG from Scratch | |
| Fintech AI Agent from Scratch | |
MultiModal
Search across different types of data (text, images, and more). Build powerful search applications that work with diverse inputs.
RAG
Generated Responses by retrieving relevant documents before answering. This section covers different approaches to implementing RAG in your projects.
Vector Search
Find relevant documents quickly! These projects show how to use vector-based search techniques to make AI-powered searches faster and smarter.
Chatbot
Create chatbots that understand user queries and fetch relevant responses using LanceDB’s vector search capabilities.
Evaluation
These projects provide tools to compare AI-generated responses against reference data and fine-tune accuracy.
| Evaluation | Interactive Notebook & Scripts | Blog |
|---|---|---|
| Monitoring and Tracing RAG using HoneyHive | | |
| Evaluating RAG with RAGAs | | |
AI Agents
Build applications where multiple AI agents interact to complete tasks efficiently. These projects show how agents can collaborate, exchange data, and automate workflows.
Recommender Systems
Personalized AI recommendations! These projects help you build recommendation engines that suggest content based on user preferences.
| Recommender Systems | Interactive Notebook & Scripts | Blog |
|---|---|---|
| Movie Recommender | | |
| Product Recommender | | |
| Arxiv paper recommender | | |
| Music Recommender | | |
Concepts
Learn the core ideas behind AI applications—including text chunking, retrieval strategies, and optimization techniques—to improve your understanding of vector search and AI pipelines.
Projects & Applications
Ready-to-use AI applications built with LanceDB! Use these projects as-is, customize them, or integrate them into your own applications.
Node applications powered by LanceDB
| Project Name | Description | Screenshot |
|---|---|---|
| Writing assistant | Writing assistant app using lanchain.js with LanceDB, allows you to get real time relevant suggestions and facts based on you written text to help you with your writing. | |
| Sentence Auto-Complete | Sentance auto complete app using lanchain.js with LanceDB, allows you to get real time relevant auto complete suggestions and facts based on you written text to help you with your writing.You can also upload your data source in the form of a pdf file.You can switch between gpt models to get faster results. | |
| Article Recommendation | Article Recommender: Explore vast data set of articles with Instant, Context-Aware Suggestions. Leveraging Advanced NLP, Vector Search, and Customizable Datasets, Our App Delivers Real-Time, Precise Article Recommendations. Perfect for Research, Content Curation, and Staying Informed. Unlock Smarter Insights with State-of-the-Art Technology in Content Retrieval and Discovery!". | |
| AI Powered Job Search | Transform your job search experience with this AI-driven application. Powered by LangChain.js, LanceDB, and advanced semantic search, it provides real-time, highly accurate job listings tailored to your preferences. Featuring customizable datasets and advanced filtering options (e.g., skills, location, job type, and salary range), this app ensures you find the right opportunities quickly and effortlessly. Best suited for job seekers, recruiters, career platforms, custom job boards. | |
| AI Powered Multimodal meme search | An advanced AI-powered meme search engine that allows users to find memes using both text and image queries. By leveraging LanceDB as a high-performance vector database and Roboflow's CLIP model for embedding generation, the platform delivers fast and accurate meme retrieval. | |
| AI Powered Feedback search and analysis | An AI-powered employee feedback analysis platform designed to collect, store, analyze, and retrieve insightful employee feedback. This system leverages LanceDB for high-speed vector-based semantic search, React.js for an interactive UI, Node.js for backend processing, and LangChain.js with an Ambient Agent for intelligent analysis and actionable insights. | |
| Hierarchical Multi Agent | The AI-Powered Law Assistant is a Hierarchical Multi-Agent System leveraging LangGraph, LangChain, and LanceDB for efficient legal query processing. It features a Supervisor Agent that delegates tasks to specialized agents for IPC and NDPS laws, each with sub-agents for case retrieval and legal summarization. Using LanceDB, it stores and retrieves vectorized legal documents, enabling fast, structured, and context-aware responses for legal professionals, researchers, and law students. | |
| Project Name | Description | Screenshot |
|---|---|---|
| YOLOExplorer | Iterate on your YOLO / CV datasets using SQL, Vector semantic search, and more within seconds | |
| Website Chatbot (Deployable Vercel Template) | Create a chatbot from the sitemap of any website/docs of your choice. Built using vectorDB serverless native javascript package. | |
| Advanced Chatbot with Parler TTS | This Chatbot app uses Lancedb Hybrid search, FTS & reranker method with Parlers TTS library. | |
| Multi-Modal Search Engine | Create a Multi-modal search engine app, to search images using both images or text | |
| Evaluate RAG | A working Streamlit RAG App designed to demonstrate end to to end production grade evaluation using 50+ scores and metrics which include guards, software metrics, traditional metrics and LLM as judge metrics. It uses mixture of specialised deep learning models and LLM as Judge models to do the evaluations | |
| Multi-Agent Collaboration Chatbot | Multi-Agent collaboration chatbot using langgraph for share-market use case using Lancedb & tools such as Polygon ,Tavily | |
| Multimodal Myntra Fashion Search Engine | This app uses OpenAI's CLIP to make a search engine that can understand and deal with both written words and pictures. | |
| Multilingual-RAG | Multilingual RAG with cohere embedding & support 100+ languages | |
| Music Recommender | Music Recommendation system using audio feature extraction and vector similarity search. By utilizing LanceDB, PANNs for audio tagging, and Librosa for audio feature extraction, the system finds and recommends tracks with similar audio characteristics based on a query song. | |
| NoOCR | End-to-end solution for complex PDFs, powered by ColPali and LanceDB. | |
🌟 New! 🌟 Applied GenAI and VectorDB course on Udacity Learn about GenAI and vectorDBs using LanceDB in the recently launched Udacity Course
Contributing Examples
If you're working on some cool applications that you'd like to add to this repo, please open a PR!
Similar repos
Contributors
(top 30 of 37)
Languages
Jupyter Notebook
92.6%
Python
3.7%
JavaScript
3.0%