UniversalRAG: Retrieval-Augmented Generation over Corpora of Diverse Modalities and Granularities

Multilingual Historical Entity Linking with Large Language Models

Source code for the FineTranslations dataset

The all-in-one Desktop & Docker AI application with built-in RAG, AI agents, No-code agent builder, MCP compatibility, and more.

Easily deploy Haystack pipelines as REST APIs and MCP Tools.

A ready-to-use docker-compose with Hayhooks and open-webui

A Redis-based Human-in-the-Loop implementation for Haystack Agents, integrated with Open WebUI for interactive tool approval workflows.

Display ZenML pipeline status on a TRMNL e-ink display

👩🏻‍🍳 A collection of example notebooks using Haystack

Bichon – A lightweight, high-performance Rust email archiver with WebUI

A corpus-free toolkit for morphological lexicon creation and tokenizer evaluation using MDL-inspired atomicity scoring for Uralic languages

This is the codebase for "Large Reasoning Models Are (Not Yet) Multilingual Latent Reasoners"

NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems.

PyTorch building blocks for the OLMo ecosystem

Introduction to Machine Learning Systems

Postgres extension for vector search (DiskANN), complements pgvector for performance and scale. Postgres OSS licensed.

Data and tools for generating and inspecting OLMo pre-training data.

Physics of Language Models, Part 4

Small Language Model Inference, Fine-Tuning and Observability. No GPU, no labeled data needed.

RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.

Universal Reasoning Model

OpenNER

Information hub for our project training the largest possible historical LLMs.

A multilingual NER dataset covering 91 languages and 25 scripts.

Code for Bolmo: Byteifying the Next Generation of Language Models