StarTrail-org/PixelRAG
Visual retrieval-augmented generation pipeline that renders web pages, PDFs and images to screenshot tiles and retrieves over them with a LoRA-tuned Qwen3-VL embedding model.
Awesome AI Agents › Document Analytics Platforms
Tencent/WeKnora
WeKnora is an advanced framework powered by large language models (LLMs) designed for deep document understanding, semantic retrieval, and context-aware question answering using the Retrieval-Augmented Generation (RAG) paradigm. It is particularly suited for handling complex and heterogeneous documents by combining multimodal preprocessing, semantic vector indexing, intelligent retrieval, and LLM inference in a modular architecture. The framework supports structured content extraction from various document formats including PDFs, Word documents, text files, markdown, and images with OCR capabilities, unifying them into semantic views for better understanding. WeKnora's architecture is modular and flexible, allowing customization and extension of components such as document parsing, embedding, retrieval, and generation. It employs hybrid retrieval strategies that combine keyword-based, vector-based, and knowledge graph-enhanced methods to ensure efficient and precise information retrieval. The system supports integration with popular vector databases like PostgreSQL (pgvector) and Elasticsearch, and embedding models can be local or cloud-based APIs. The framework offers intelligent reasoning capabilities by leveraging LLMs to understand document context and user intent, enabling accurate multi-turn conversations and Q&A. It provides a user-friendly web interface and standardized RESTful APIs, making it accessible to both developers and business users without technical barriers. Security is emphasized with login authentication and recommendations for private network deployment to protect data sovereignty. WeKnora is applicable in various domains such as enterprise knowledge management, academic research, product technical support, legal and compliance review, and medical knowledge assistance. It supports end-to-end testing with visualization and metric evaluation for retrieval and generation processes. Deployment options include local and Docker-based setups, catering to private and offline environments. Overall, WeKnora is a comprehensive, secure, and extensible framework that facilitates deep semantic understanding and retrieval of documents, enhancing knowledge discovery, decision-making, and customer support across multiple industries.
https://github.com/Tencent/WeKnora
Also in Document Analytics Platforms
georgia-tech-db/evadb
EvaDB is an open-source database system that empowers developers to build AI-powered applications by integrating advanced AI models and optimizations within a SQL-based querying framework for both structured and unstructured data.
loggie-io/loggie
Loggie is a lightweight, cloud-native data transfer agent and aggregator designed for scalable, high-performance log collection, processing, and monitoring in Kubernetes environments.
johnsonhk88/AI-Bank-Statement-Document-Automation-By-LLM-And-Personal-Finanical-Analysis-Prediction
A full-stack pipeline that parses bank statement PDFs with layout detection, OCR and LLMs, then runs CrewAI, LangChain Deep Agents and Hermes agent harnesses over the extracted data for categorisation, analysis and RAG querying.
netboxlabs/pktvisor
pktvisor is a dynamic network observability agent that analyzes high volume network traffic and generates opentelemetry metrics for real-time and centralized monitoring.
genomoncology/FuzzTypes
FuzzTypes is a Pydantic extension that provides autocorrecting annotation types for enhanced data validation and normalization, enabling automatic correction and structured data composition in Python applications.