Builds retrieval-augmented generation pipelines from natural language instructions. Defaults to OpenAI models, requires external API access, and runs locally as a self-hosted web GUI.
Project overview
Builds retrieval-augmented generation pipelines from natural language instructions. Defaults to OpenAI models, requires external API access, and runs locally as a self-hosted web GUI.
Project type
AI Agent · RAG · Workflow
Use cases
Knowledge Q&A
Deployment
Refer to project documentation
License
MIT
Best for
Developers and general users who want to generate a RAG pipeline using natural language and have access to external model APIs like OpenAI.
Key capabilities
Instruct a builder agent with natural language descriptions of your dataset, task, and parameters to construct the pipeline.
View and manually edit generated RAG parameters such as top-k, chunk size, embed model, LLM, and system prompt.
Query the generated RAG agent and receive answers over your data via a chatbot interface.
Customize the builder agent to use different LLMs, defaulting to OpenAI with support for Anthropic.
Configure the generated RAG agent to use OpenAI, Anthropic, Replicate, or HuggingFace LLMs, and OpenAI or HuggingFace embeddings.
Limitations and risks
Claude models currently do not work reliably for constructing the agent; GPT-4 variants are recommended.
Adoption risk: upgrading the version may introduce breaking changes in the stored data structure, requiring cache deletion.
Getting started
Clone the project, create a virtual environment, run poetry install --with dev, add .streamlit/secrets.toml with your openai_key, and run streamlit run 1_🏠_Home.py.
Evidence and sources
README: RAGs is a Streamlit app that lets you create a RAG pipeline from a data source using natural language.
README: Describe your task (e.g. "load this web page") and the parameters you want from your RAG systems (e.g. "i want to retrieve X number of docs")
README: By default, we use OpenAI for both the builder agent as well as the generated RAG agent. Add `.streamlit/secrets.toml` in the home folder. Then put the following: openai_key = "<o…
README: Currently we support either **a single local file** or a **web page**.
README: In this section, you have a UI showcasing the generated parameters and have full freedom to manually edit/change them as necessary.