Awesome AI Agents › Generative Software Frameworks

deepset-ai/haystack

⭐ 26502 Python added to this list on 2025-04-09 repository created 2019-11-14

Haystack by deepset-ai is an advanced AI orchestration framework designed to build customizable, production-ready applications powered by large language models (LLMs). It enables users to connect various components such as models, vector databases, and file converters into pipelines or agents that can interact with data effectively. The framework is particularly suited for building retrieval-augmented generation (RAG) systems, question answering applications, semantic search engines, and conversational agent chatbots. Haystack supports a wide range of technologies and vendors, allowing users to choose from models available from OpenAI, Cohere, Hugging Face, or even local and cloud-hosted models on platforms like Azure, Bedrock, and SageMaker. This technology-agnostic approach provides flexibility and transparency in how components communicate, making it easier to integrate with existing tech stacks and tailor solutions to specific use cases. The framework offers comprehensive tooling in one place, including database access, file conversion, data cleaning, splitting, training, evaluation, and inference. It is designed to be extensible, encouraging community and third-party contributions to build custom components and foster an open ecosystem. Haystack supports scaling to millions of documents using retrievers and production-scale components, and it allows users to fine-tune models on their own data or use off-the-shelf models. Additionally, it incorporates user feedback mechanisms to evaluate, benchmark, and continuously improve model performance. Haystack is accessible via pip installation and also supports Docker images for deployment. Extensive documentation, tutorials, and a cookbook of recipes are available to help users get started and explore advanced use cases. Overall, Haystack is a powerful and flexible framework for developing sophisticated NLP applications that leverage the latest advancements in LLMs and vector search technologies.

https://github.com/deepset-ai/haystack

agentsaiai-orchestration-frameworkazurebedrockbertchatbotschatgptcohereconversational-agentscustomizable-pipelinesdata-cleaningdatabase-accessevaluationextensiblefile-conversionfile-convertersfine-tuningflexiblegenerative-aihugging-faceinferenceinformation-retrievallanguage-modellarge-language-modelsllmllm-applicationsmachine-learningnlpnlp-applicationsopenaiproduction-readypythonpytorchquestion-answeringragretrieval-augmented-generationsagemakersemantic-searchsquadsummarizationtechnology-agnostictrainingtransformersuser-feedbackvector-databasesvector-search