RAG Frameworks, Parsers and Vector Databases

ZettaBrain RAG is a self-hosted document assistant for individuals and small teams working with policies, contracts and internal manuals. It helps them find answers and draft documents from files they already keep on their own infrastructure.

OpenVINO Model Server is a self-hosted inference server for developers and teams serving AI models on their own machines or servers. It uses OpenVINO to run workloads on x86 CPUs, including Xeon, with acceleration for Intel integrated and discrete GPUs and NPUs. It's open source under Apache 2.0.

Dolphin Document Parser converts PDFs and document images into structured content for developers working on document extraction. It handles digitally created and photographed pages, including documents that mix text with tables, formulas, and code.

Codanna helps developers and AI coding agents explore a repository and assess which code a change could affect. It connects code search with function relationships, so an assistant can get context around a match in one response.

MDFlux is a desktop app for turning documents into structured Markdown for AI workflows, research, or text editing. It suits people who need usable text from mixed files and scanned archives while keeping conversion on their own machine.

Memvid gives AI agents persistent memory in a portable file. It's for developers who need agents to recall conversations or search documents without running a separate database server.

LLPhant is a PHP framework for developers adding AI capabilities to Symfony and Laravel applications. It supports local model services and cloud providers, so you can choose where model requests run.

pgvectorscale extends pgvector for developers who want to search large embedding datasets within PostgreSQL. It adds indexing and compression for AI applications that keep vector data in their existing database.

Vespa is an AI search platform for developers building search, RAG and recommendation applications over large datasets that change continuously. It retrieves candidate results and evaluates their relevance with machine-learned models within the same platform.

Bisheng is a self-hosted AI application platform for teams working with business documents and processes. It supports tasks such as document review, support ticket assistance and fixed-layout report generation, including workflows that need human judgment during execution.

Open Notebook is an alternative to Google's NotebookLM for researchers, students and professionals working with their own materials. It brings research and notes together while letting you choose what context an AI model can access.

EverOS is a Python library and memory runtime for developers building agents that need context across sessions and applications. It keeps conversations, files and task histories in readable Markdown that you can edit and version with Git.

PipesHub connects company knowledge to AI search and agents for teams whose files and conversations sit across business apps. It can also supply that context to coding assistants and custom applications.

M Flow is a self-hosted memory engine for developers whose AI agents need to recall past conversations and work preferences. It runs in Python or Docker and supports Ollama for local model use, alongside providers such as OpenAI and Anthropic. Choosing a cloud provider sends model requests to that service.

LiteRT-LM is Google's open-source framework for developers building apps that run language models on the user's device. It uses LiteRT for local inference across mobile devices, desktops and browsers, with GPU and NPU acceleration. The project uses the Apache 2.0 license.

MindsDB Query Engine gives developers building AI agents one SQL interface for business data and document search. It runs as a standalone server and can also provide data access to MindsHub agents.

Neo4j LLM Graph Builder converts documents and web content into knowledge graphs for people building searchable collections or AI applications with Neo4j. It uses an LLM to identify entities, relationships, and their properties.

MinerU is a local document parser for people preparing files for AI workflows. It handles scanned PDFs, academic papers and other documents containing OCR text, tables, figures or complex layouts.

BGE Embeddings is a family of models for developers who need to find relevant passages for search or retrieval-augmented generation (RAG). It includes embedding models for finding candidate text and rerankers for assessing how closely those results match a query.

TaxHacker is a self-hosted AI accounting app for freelancers, indie hackers, and small businesses that need to turn receipts and invoices into searchable transactions. The project is in early development. Its open source edition runs on your own infrastructure or home server through Docker and uses the MIT license.

dots.ocr is a self-hosted OCR model for developers and teams that need text and page structure from multilingual documents. It combines text recognition and layout parsing in a single vision-language model you can run on your own hardware.

Quivr is a content search and monitoring engine for developers building applications around incoming material and topics they want to follow over time. It's at an evaluation stage, and the project advises against production use because its API may change without notice.

Gitingest prepares code context for developers who want to give an LLM a repository or local directory. It gathers the project structure and file contents into a text digest instead of requiring you to assemble files by hand.

Dify is a visual builder for teams creating AI agents and applications that use their own data and tools. Its canvas makes application logic visible so teams can inspect and test how a task proceeds.

sage-wiki turns a document collection into a linked wiki and knowledge graph for researchers, teams, and AI agents that need reusable knowledge across tasks.

QAnything is a local knowledge base for people and teams who want to ask questions about their own documents. It supports offline operation and questions across Chinese and English document collections.

ExtractThinker is a Python library for developers who want to extract structured data from documents using models running locally through Ollama. It also supports OpenAI, Anthropic and Groq. With a cloud provider, extraction sends document content to that service, requires provider credentials and may incur charges.

LanceDB is a vector database for developers building AI retrieval applications and teams working with multimodal datasets. Its embedded library can run locally or in your own cloud.

Infinity Embeddings is a self-hosted model server for developers building semantic search and retrieval-augmented generation applications. It runs retrieval models on your own hardware and can serve several models together.

Marqo Open Source is a self-hosted search engine for developers building document search, product discovery, or retrieval for AI applications. The project is deprecated and no longer receives updates.

llmware is a Python framework for developers building applications that answer questions using their own documents or carry out multi-step AI agent workflows. It focuses on small, specialized models that can run on laptops, edge devices, and self-hosted servers.

PageIndex is a document retrieval engine for developers building question answering over long PDFs, including financial reports, legal documents and technical manuals. It uses an LLM to select relevant sections from a document tree without a vector database or document chunking.

GraphRAG is an open-source Python indexing and retrieval pipeline for developers building question answering over their own document collections. It can connect to local models through Ollama proxies, while OpenAI models are its default supported choice. With a remote model provider, text sent for model processing leaves your machine.

USearch is a vector similarity search library for developers building semantic search and recommendation systems within their own apps or servers. Its compact C++ core supports custom similarity rules and has no mandatory dependencies.