Environment:Deepset ai Haystack Python Runtime Environment

Knowledge Sources	Haystack pyproject.toml
Domains	Infrastructure, NLP
Last Updated	2026-02-11 20:00 GMT

Overview

Python 3.10+ runtime environment with core dependencies for building and running Haystack LLM pipelines.

Description

This environment defines the base Python runtime and core package dependencies required to run any Haystack pipeline. It provides the foundation for the framework including pipeline graph orchestration via NetworkX, template rendering via Jinja2, OpenAI API integration, serialization, and structured logging. All Haystack components inherit this base environment. The framework is OS-independent and supports Linux, macOS, and Windows.

Usage

Use this environment for any Haystack pipeline or component execution. This is the mandatory base prerequisite for all Implementations in the Haystack framework, from document indexing to RAG pipelines to evaluation workflows.

System Requirements

Category	Requirement	Notes
OS	OS Independent	Linux, macOS, Windows all supported
Python	>= 3.10	Officially tested: 3.10, 3.11, 3.12, 3.13
Disk	Minimal	Core package is lightweight; models require additional space

Dependencies

System Packages

No special system packages required for base functionality

Python Packages

`haystack-ai` (the framework itself)
`openai` >= 1.99.2
`typing_extensions` >= 4.7
`networkx` (pipeline DAG orchestration)
`Jinja2` (prompt template rendering)
`pydantic` (data validation)
`pyyaml` (serialization)
`numpy` (numerical operations)
`tqdm` (progress bars)
`tenacity` != 8.4.0 (retry logic; 8.4.0 excluded due to bugs)
`lazy-imports` (optional dependency management)
`posthog` != 3.12.0 (telemetry; 3.12.0 excluded due to bugs)
`more-itertools` (used by TextDocumentSplitter)
`requests` (HTTP client)
`python-dateutil` (date parsing)
`jsonschema` (JSON schema validation for Tools)
`docstring-parser` (ComponentTool introspection)
`filetype` (MIME type detection)
`haystack-experimental` (experimental features)

Credentials

No credentials are required for the base runtime. Individual components (OpenAI generators, HuggingFace embedders) require their own API tokens. See:

Environment:Deepset_ai_Haystack_OpenAI_API_Environment for OpenAI API keys
Environment:Deepset_ai_Haystack_HuggingFace_Model_Environment for HuggingFace tokens

Quick Install

pip install haystack-ai

Code Evidence

Python version requirement from `pyproject.toml:11`:

requires-python = ">=3.10"

Core dependencies from `pyproject.toml:43-62`:

dependencies = [
  "tqdm",
  "tenacity!=8.4.0",
  "lazy-imports",
  "openai>=1.99.2",
  "pydantic",
  "Jinja2",
  "posthog!=3.12.0",
  "pyyaml",
  "more-itertools",
  "networkx",
  "typing_extensions>=4.7",
  "requests",
  "numpy",
  "python-dateutil",
  "jsonschema",
  "docstring-parser",
  "filetype",
  "haystack-experimental",
]

Version detection from `haystack/version.py:5-10`:

from importlib import metadata

try:
    __version__ = str(metadata.version("haystack-ai"))
except metadata.PackageNotFoundError:
    __version__ = "main"

Common Errors

Error Message	Cause	Solution
`ModuleNotFoundError: No module named 'haystack'`	Package not installed	`pip install haystack-ai` (not `pip install haystack`)
`ImportError: tenacity`	Incompatible tenacity version 8.4.0	`pip install "tenacity!=8.4.0"`
`ImportError: posthog`	Incompatible posthog version 3.12.0	`pip install "posthog!=3.12.0"`

Compatibility Notes

Package name: Install `haystack-ai`, not `haystack` (the old Haystack 1.x package)
Python 3.9 and below: Not supported; requires 3.10+ for typing features
tenacity 8.4.0: Explicitly excluded due to known issues
posthog 3.12.0: Explicitly excluded due to known issues (https://github.com/PostHog/posthog-python/issues/187)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment