Repolode
Submit
deepset-ai

deepset-ai/haystack

Open-source Python framework for building production-ready LLM applications with modular pipelines and agent workflows for RAG, search and agents.

View on GitHubWebsite
Stars
26,786
Forks
3,307
Open issues
149
Last commit
8 hours ago

What is it?

Haystack is an open-source AI orchestration framework written in Python. It lets developers assemble LLM applications from modular pipelines and agent workflows, with explicit control over how documents are retrieved, ranked, routed, stored in memory and passed to a model.

The framework targets retrieval-augmented generation, semantic search, question answering, multimodal applications and conversational systems. Pipelines can run synchronously or asynchronously and stream tokens, and agents support lifecycle hooks, tool calls and built-in tracking of step count and token usage.

Haystack works with many model providers, including OpenAI, Mistral, Anthropic, Cohere, Hugging Face, Google, Azure OpenAI and AWS Bedrock, as well as local models. Hayhooks, a separate project, can expose pipelines as REST endpoints or MCP servers.

Key features

  • 01Modular pipelines with loops, branches and conditional logic for routing context.
  • 02Agents with lifecycle hooks, tool calling and built-in step and token usage tracking.
  • 03Synchronous and asynchronous pipeline execution with token streaming.
  • 04Integrations with major LLM providers and local models, swappable without rewriting the system.
  • 05Consistent component interface for building and sharing custom components.

Use cases

  • Building retrieval-augmented generation systems over private document collections.
  • Creating semantic search or question answering over enterprise content.
  • Developing autonomous or multi-agent workflows with tool use.
  • Deploying LLM pipelines as HTTP or MCP services using Hayhooks.

Quick start

  1. 1

    Install Haystack via pip

    pip install haystack-ai
  2. 2

    Optionally install nightly pre-releases to try the newest features

    pip install --pre haystack-ai
  3. 3

    Read the Get Started Guide in the documentation to build your first LLM application

Things to know

  • Some integrations, such as model vendors, require separate API keys or accounts.

Topics

#llm#rag#agents#semantic-search#orchestration#python#retrieval-augmented-generation#agent-framework#agentic-ai#agentic-rag

FAQ

What is haystack?
Open-source Python framework for building production-ready LLM applications with modular pipelines and agent workflows for RAG, search and agents. Haystack is an open-source AI orchestration framework written in Python. It lets developers assemble LLM applications from modular pipelines and agent workflows, with explicit control over how documents are retrieved, ranked, routed, stored in memory and passed to a model.
Is haystack free to use? What license does it have?
haystack is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is haystack still maintained?
Actively maintained. The last commit to haystack was on Oct 20, 2026.
What can I use haystack for?
Building retrieval-augmented generation systems over private document collections. Creating semantic search or question answering over enterprise content. Developing autonomous or multi-agent workflows with tool use. Deploying LLM pipelines as HTTP or MCP services using Hayhooks.
How do I install haystack?
Install Haystack via pip: pip install haystack-ai Optionally install nightly pre-releases to try the newest features: pip install --pre haystack-ai Read the Get Started Guide in the documentation to build your first LLM application
Where can I get haystack?
The source code and installation instructions are on GitHub: https://github.com/deepset-ai/haystack

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects