Llamafarm

LlamaFarm is an open source platform, released under the Apache 2.0 license, for deploying AI models and the supporting pieces around them, including databases, retrieval augmented generation pipelines, agents and other pipelines. It is built to run locally on the user's own device or in a private environment, so that applications keep working when a cloud provider is down or throttling, and it can also be deployed to the cloud when scale is needed. It is aimed at developers who want to move from prototype to production using a command line first workflow and JSON or YAML configuration rather than extensive code. Projects are initialized and run with the lf command line tool. The platform covers multiple databases, parsing strategies, embedding models and retrieval strategies for RAG, and exposes an OpenAI compatible API so that an existing client can be pointed at a project with its own knowledge base attached. Newer features include an AI assisted designer for building projects locally, a Universal Runtime for running transformer models on a laptop or in the cloud, and built in speech to text and text to speech testing with Whisper and Piper TTS. Beyond language models, the site lists support for anomaly detection, text classifiers based on models such as BERT and RoBERTa, time series forecasting and vision models for tasks like OCR, image classification and object detection. The software is free and open source, with its code on GitHub. The site does not publish any paid plans or prices.