AI Agents & Automation

obsidian-Smart2Brain is a free and open-source Obsidian plugin designed to transform your second brain into a smarter, more interactive knowledge base. It functions as a personal AI assistant, powered by large language models such as ChatGPT or Llama2. A key feature is its ability to directly access and process your notes, eliminating the need for manual prompt editing. For privacy-conscious users, it can operate completely offline, ensuring your data remains secure and private. The plugin utilizes a RAG pipeline to embed notes into vectors, retrieving relevant information based on your query to generate answers with reference links to the original notes. Users can choose between different chat views and easily switch between various LLMs, including local models via Ollama, or leverage OpenAI's advanced capabilities if preferred. It also allows saving and continuing chat conversations.

NeumAI is a robust data platform designed to empower developers in leveraging their data for contextualizing Large Language Models (LLMs) through Retrieval Augmented Generation (RAG). It streamlines the process of extracting data from various sources, including document storage and NoSQL databases, processing this content into vector embeddings, and then ingesting these embeddings into vector databases for efficient similarity search. The platform offers a high-throughput, distributed architecture capable of handling billions of data points, ensuring optimal parallelization for embedding generation and ingestion. Key features include built-in connectors for common data sources, embedding services, and vector stores, along with real-time data synchronization. NeumAI also provides customizable data pre-processing options and cohesive data management to support hybrid retrieval with augmented metadata, reducing the time spent on integrating diverse services.

RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that integrates cutting-edge RAG with Agent capabilities to create a superior context layer for Large Language Models (LLMs). It offers a streamlined RAG workflow adaptable for enterprises of any scale, powered by a converged context engine and pre-built agent templates. RAGFlow enables developers to transform complex data into high-fidelity, production-ready AI systems with exceptional efficiency and precision. Key features include deep document understanding, template-based chunking, grounded citations with reduced hallucinations, compatibility with heterogeneous data sources, and an automated RAG workflow. It supports various LLMs and embedding models, multiple recall, and fused re-ranking.

EnFi, Inc is an AI-powered platform designed to transform lending operations for banks, private credit firms, credit unions, and FinTech lenders. It tackles document chaos by ingesting and processing various financial documents, extracting meaningful insights, and driving actionable lending decisions. The platform offers solutions for deal screening and underwriting, converting unstructured documents into deal analyses and providing clear term recommendations. It also excels in portfolio monitoring and risk analysis, turning borrower reporting into risk assessments and offering intervention recommendations. EnFi's multi-agent architecture utilizes specialized AI agents and domain-specific models trained on financial documents, ensuring explainable AI with full decision traceability and enterprise-grade security.

textbook_quality is an open-source project designed to generate high-quality synthetic pretraining data for large language models (LLMs). It offers robust capabilities for creating extensive datasets, exemplified by its ability to produce 70M token examples. The tool supports parallel generations, allowing users to leverage OpenAI or their own custom APIs for data creation. A key feature is its use of retrieval mechanisms, such as Serply or SerpAPI, to significantly improve the quality of the generated content. Users can either generate topics from scratch based on a subject and desired iterations or augment existing seed topics semantically. The core architecture is extensible, enabling developers to integrate new LLM adaptors, retrieval methods, and tasks, making it a flexible solution for advanced LLM data generation.

vlite is a simple and blazing fast vector database built entirely in NumPy, eliminating the need for complex database setups or running servers. It's specifically designed for Retrieval-Augmented Generation (RAG) applications, offering exceptionally fast vector retrieval using binary embeddings. A key innovation is its CTX file format, which acts like a browser cookie for user embeddings, ensuring efficient storage, retrieval, portability, and user context. vlite supports ingesting various data types including text, PDF, CSV, PPTX, and webpages, with features like chunking, metadata filtering, and PDF OCR for scanned documents. Benchmarks show it to be significantly faster than alternatives like Chroma on indexing and retrieval, while also consuming less disk space. It is available as a Python package and integrates seamlessly with LangChain.

Lili.ai leverages AI and Natural Language Processing (NLP) to analyze vast amounts of textual data from project documentation, such as meeting minutes, correspondence, and daily reports. This allows project managers to unlock hidden insights, detect potential problems, and make informed decisions at an early stage, significantly reducing the risk of cost overruns, litigation, and project stalemates. The system is specifically trained on the language of major projects, enabling it to accurately identify and track risk mentions like delays, overcosts, and quality deficiencies. It also assists in quickly finding written mentions of problems for gap analysis or claim memorandum writing, saving valuable time for project teams.

AgentSearch is a robust framework designed to empower search agents by integrating Large Language Model (LLM) technologies from diverse providers with various search engines. This integration facilitates Retrieval-Augmented Generation (RAG), enabling search agents to perform a wide array of functions, including summarizing search results, generating new queries, and retrieving detailed downstream results. Key features include effortless search agent building by connecting search-specialized LLMs (like Sensei-7B) with supported search engines, customizable local search using the AgentSearch dataset, and seamless API endpoint integration with hosted providers such as Bing, SERP API, and AgentSearch itself. It also supports LLMs from SciPhi, HuggingFace, OpenAI, and Anthropic, offering flexibility and ease of use for developers.

Rebuzz AI is an AI-powered content calendar designed to streamline social media marketing by generating 30 days of brand-consistent social media posts in under 10 minutes. Leveraging RAG (Retrieval-Augmented Generation) technology, the tool learns and adapts to your specific brand voice, ensuring all generated content aligns with your established identity. This makes it an efficient solution for businesses and marketers looking to maintain a consistent online presence without extensive manual effort. It offers a free trial, allowing users to experience its capabilities in automating content creation and calendar management.

BloombergGPT is a 50-billion parameter large language model developed by Bloomberg, specifically designed for the financial industry. It has been trained from scratch on a vast dataset of financial data to support a diverse range of natural language processing (NLP) tasks. This model demonstrates superior performance on financial tasks compared to other open models of similar size, while also maintaining or exceeding their performance on general NLP benchmarks. Its specialized training makes it highly effective for applications within finance, offering advanced capabilities for analyzing and interpreting complex financial information.

Cognita is an open-source RAG (Retrieval Augmented Generation) framework designed to streamline the development and deployment of modular, scalable, and extensible AI applications. While leveraging Langchain/LlamaIndex, Cognita addresses the challenges of moving RAG systems from experimentation to production by offering an organized codebase where each RAG component is API-driven. It supports various features like multiple document retrievers, incremental indexing, and integration with open-source LLMs and embedding models. Cognita also includes a no-code UI for easier configuration and experimentation, making it suitable for both local development and production environments, with optional support for Truefoundry components for enhanced scalability.

Summarify.me is an AI-powered summarization tool designed to quickly convert long-form content into concise summaries. It supports a wide range of input types, including text, web URLs, blog URLs, YouTube videos, PDFs, and audio files. Users can paste text directly, upload documents, or provide links to generate detailed summaries in seconds. The tool offers various pricing plans, including a free trial, with increasing limits on the number of summaries and file sizes for PDFs, YouTube videos, and audio files. Summarify.me aims to help users save time by providing instant distillations of information, making it ideal for students, researchers, and professionals who need to quickly grasp the main points of extensive content.

Cheshire Cat AI, also known as core, is an open-source framework designed for building custom AI agents as a microservice. It adopts an API-first philosophy, making it easy to integrate conversational layers into existing applications. Key features include chat via WebSocket, a customizable REST API for agent management, and built-in Retrieval Augmented Generation (RAG) using Qdrant. The framework is highly extensible through plugins, supports event callbacks, function calling (tools), and conversational forms. It also provides an easy-to-use admin panel, supports any language model via Langchain, and offers multi-user capabilities with granular permissions. The entire system is 100% dockerized, ensuring easy deployment and scalability. It also boasts an active Discord community and comprehensive documentation.

Korvus is an all-in-one, open-source RAG (Retrieval-Augmented Generation) pipeline built specifically for Postgres. It integrates LLMs, vector memory, embedding generation, reranking, summarization, and custom models into a single SQL query, significantly boosting performance and simplifying search architecture. By leveraging Postgres's robust capabilities, Korvus eliminates the need for external services and API calls, reducing latency and complexity. It provides SDK support for Python, JavaScript, Rust, and C, allowing seamless integration into existing tech stacks. This approach offers a simplified architecture, high performance, and scalability, making it ideal for developers looking to build efficient RAG applications directly within their database.

llm-search offers an advanced Retrieval-Augmented Generation (RAG) system designed for querying local documents using Large Language Models. It features a simple YAML-based configuration for easy setup and interaction. The tool focuses on enhancing various RAG components, including improved document parsing for formats like Markdown, PDF, and DOCX, hybrid search capabilities, and re-ranking strategies. It supports custom LLMs, including OpenAI compatible models, HuggingFace models, and interoperability with LiteLLM + Ollama. Users can interact with the system via a built-in frontend or an MCP server, allowing integration with clients like Cursor, Windsurf, or VSCode GH Copilot. Additional features include incremental updates for document bases, chat history, and optional HyDE and multi-querying for improved result quality.

LangChain4j is an idiomatic, open-source Java library designed to streamline the integration of Large Language Models (LLMs) into Java applications running on the JVM. It offers a unified API that abstracts away the complexities of various LLM providers (like OpenAI, Google Vertex AI) and embedding stores (such as Pinecone, Milvus), allowing developers to easily switch between them without extensive code changes. The library provides a comprehensive toolbox for common LLM patterns and techniques, including low-level prompt templating, chat memory management, function calling, and high-level patterns like Agents and Retrieval Augmented Generation (RAG). LangChain4j is built with Java conventions in mind, emphasizing type safety, POJOs, annotations, and seamless integration with enterprise Java frameworks like Quarkus and Spring Boot. It aims to provide a robust and developer-friendly solution for creating sophisticated LLM-powered applications in Java.

LUKE (Language Understanding with Knowledge-based Embeddings) is a cutting-edge pretrained contextualized representation model for words and entities, built upon a transformer architecture. It incorporates entity-aware self-attention to achieve state-of-the-art performance across various NLP benchmarks, including SQuAD v1.1 for extractive question answering, CoNLL-2003 for named entity recognition, ReCoRD for cloze-style question answering, TACRED for relation classification, and Open Entity for entity typing. The repository provides source code for both pretraining the model and fine-tuning it for downstream tasks. LUKE models, including Japanese versions and lite versions with reduced memory footprint, are readily available on the Hugging Face Model Hub, making them accessible for researchers and developers.

Milvus is a high-performance, cloud-native vector database designed for scalable Approximate Nearest Neighbor (ANN) search. Written in Go and C++, it leverages hardware acceleration for CPU/GPU to achieve best-in-class vector search performance. Its fully-distributed and K8s-native architecture allows horizontal scaling to handle tens of thousands of search queries on billions of vectors, with real-time streaming updates. Milvus supports various vector index types, including HNSW, IVF, FLAT, SCANN, and DiskANN, and offers advanced features like metadata filtering and range search. It also supports sparse vectors for full-text search and hybrid search, combining semantic and full-text capabilities. Milvus ensures data security through user authentication, TLS encryption, and Role-Based Access Control (RBAC), making it suitable for enterprise applications.

Paper2Code is an innovative tool designed to automate the generation of code from scientific papers, specifically in the field of machine learning. At its core is PaperCoder, a sophisticated multi-agent LLM system that streamlines the process of converting research papers into functional code repositories. This system operates through a meticulously structured three-stage pipeline: planning, analysis, and code generation, with each stage handled by specialized AI agents. Paper2Code aims to significantly reduce the manual effort and time typically required to implement research findings, offering a robust solution for developers and researchers. It supports both OpenAI API and open-source models like DeepSeek-Coder-V2-Lite-Instruct, providing flexibility in deployment and cost management. The tool also includes comprehensive benchmark datasets and a model-based evaluation system for assessing the quality of generated repositories.

paperless-gpt enhances document digitalization by integrating Large Language Models (LLMs) and LLM Vision (OCR) with paperless-ngx. This tool automates the generation of document titles, tags, and correspondents, significantly reducing manual sorting time. It offers advanced OCR capabilities, leveraging models like OpenAI's GPT-4o or Ollama's qwen3:8b, to accurately extract text from even messy or low-quality scans. Key features include automatic custom field generation, searchable and selectable PDF creation, and extensive customization options for AI prompts. It supports various OCR providers including LLM-based OCR, Google Document AI, Azure Document Intelligence, and Docling Server, making it a versatile solution for efficient document organization.

SGPT (GPT Sentence Embeddings for Semantic Search) is an open-source tool that provides code, results, and pre-trained models for applying GPT models as Bi-Encoders or Cross-Encoders for symmetric or asymmetric search. SGPT-BE generates semantically meaningful sentence embeddings through contrastive fine-tuning of bias tensors and position-weighted mean pooling. SGPT-CE utilizes log probabilities from GPT models without requiring any fine-tuning. The project offers easy integration with Sentence Transformers and provides multilingual BLOOM SGPT models. Recent updates include GRIT & GritLM, which unify various SGPT functionalities into single, higher-performing models, and improved 5.8B Bi-Encoder models with better performance on USEB and BEIR benchmarks.

GPT Book Club is an AI-driven search tool designed to help users quickly grasp the essence of books. It extracts key insights and summaries, providing minified book versions and highlighting relevant passages. The platform aims to offer interactive reading experiences and resources, such as Notion templates, to enhance understanding. It is ideal for individuals who want to efficiently consume book content and identify critical information without reading entire texts, making it a valuable resource for students, researchers, and busy professionals.

TrustGraph is a comprehensive context development platform designed for building intelligent AI applications. It provides graph-native infrastructure for storing, enriching, and retrieving structured knowledge at scale, functioning like a Supabase built around context graphs. The platform features multi-model and multimodal database capabilities, automated data ingest with semantic similarity retrieval, and out-of-the-box RAG pipelines including DocumentRAG, GraphRAG, and OntologyRAG. It supports both single and multi-agent systems, integrates with MCP, and can be deployed locally with Docker or in the cloud with Kubernetes. TrustGraph offers extensive API support for major LLMs and includes a full developer toolkit with REST, Websocket, and Python APIs, along with a CLI. A key differentiator is its focus on portable, versioned Context Cores, which bundle structured knowledge, embeddings, evidence, and policies into reusable artifacts.

Verba, also known as The Golden RAGtriever, is a community-driven open-source application designed to offer a streamlined and user-friendly interface for Retrieval-Augmented Generation (RAG). It enables users to easily explore datasets and extract insights, supporting both local deployments with Ollama and Huggingface, as well as LLM providers like Anthropic, Cohere, and OpenAI. Verba is fully customizable, acting as a personal assistant for querying and interacting with data, whether locally or via cloud deployment. It combines state-of-the-art RAG techniques with Weaviate's context-aware database, offering choices between various RAG frameworks, data types, chunking and retrieving techniques, and LLM providers to suit individual use-cases. The project encourages community contributions to enhance its features and maintenance.