Description
RLAMA is a command-line platform for building retrieval-augmented generation systems and AI agents with local models. You point it at a folder of documents, such as text, Markdown and PDF files, and it creates a RAG system that answers questions from those documents, with semantic chunking and embeddings generated on your machine. Its agents can be given a role, such as researcher, writer, coder or analyst, and tools for RAG search, code execution and web search. Several agents can be grouped into a crew that works through tasks together in sequence or in parallel. The site also describes a visual RAG builder for creating a system in about two minutes, and an interactive session mode. The platform stores data locally and sends nothing to external services, which the site presents as a privacy benefit. Install commands are provided for macOS, Linux and Windows, and the site shows examples such as creating a documentation RAG and a research crew. The site notes that the project is currently paused, because the developer is busy with full-time work and university studies, and that development will resume when the schedule allows. The source is on GitHub.…
Highlights
- RAG from your own documents
- AI agents with roles and tools
- Crews of agents for tasks
- Runs with local models
- Visual RAG builder
Ratings & Reviews
No ratings yet
Be the first to rate RLAMA.
People also visit
RLAMA FAQ
What is RLAMA?
A platform for building retrieval-augmented generation systems and AI agents that run with local models.
Does my data leave my machine?
The site says RLAMA stores and processes data locally, with no data sent externally.
Is the project still active?
The site says development is paused while the developer is busy with work and studies, and that it will resume when possible.
What document types does it read?
Formats such as .txt, .md and .pdf, according to the site.
What are the best RLAMA alternatives?
Popular alternatives to RLAMA include Langflow, LocalAGI and Dialog. See all RLAMA alternatives




