$ show deepset-ai/haystack-core-integrations
This is a monorepo of 80+ Haystack integrations maintained by deepset, extending the Haystack framework with production-ready components for RAG, search, and agent pipelines. Each integration is a Python package wrapping an external service (AWS, OpenAI, Cohere, Jina, Brave, MongoDB, Qdrant, Azure, etc.) or providing specialized NLP/agent functionality. The code is library code, not research code — every component follows the Haystack component contract: @component decorator, default_to_dict/from_dict serialization, lazy warm_up, run/run_async, close_async, and Secret-based authentication. The dominant architecture is a collection of thin-to-medium wrappers unified by Haystack's pipeline abstractions, with consistent FilterPolicy (REPLACE/MERGE) document filtering and output type declarations across all integrations.
The LLM identified 4 files worth highlighting for reuse:
Deep research agent wiring multiple LLMs and search tools into a Haystack pipeline, plus advanced-RAG agent hooks and a GitHub issue-assistant toolset. Demonstrates non-obvious orchestration of heterogeneous model calls, tool use, and memory (Mem0 long-term memory example). Reusable as a reference architecture for building compound agent pipelines on Haystack.
TransformersZeroShotDocumentClassifier and related components implement the full Haystack component lifecycle: __init__, to_dict/from_dict, warm_up for lazy model loading, and run with input validation and output formatting. The test file demonstrates the canonical component pattern used across the repo. Extractable as a template for wrapping any HuggingFace model as a Haystack component.
OpenAPI integration that generates Haystack components from OpenAPI service definitions, enabling LLM agents to call arbitrary REST APIs. Test fixtures include complex type schemas (complex_types_openapi_service.json) showing non-trivial request/response handling. Reusable pattern for turning any OpenAPI-specified service into a tool-calling component.
Integration scaffolding script that generates new integration packages, updates .github/labeler.yml, README.md, and CI workflows. Demonstrates a repeatable, convention-enforcing codegen pattern that keeps 80+ integrations structurally consistent. Extractable for any monorepo needing standardized package generation.