This project has been archived by its maintainers, and is no longer receiving any updates.
Haystack experimental package
[!WARNING] This project is archived and no longer maintained.
0.19.0.post1is the final release. There will be no further releases, bug fixes, or compatibility updates, and the repository is read-only.
Most experiments graduated into Haystack itself. If you are looking for
Agent,Tooland tool calling,AsyncPipeline, pipeline breakpoints, multimodality,SuperComponent,QueryExpander, Human-in-the-Loop and more, they all ship inhaystack-ainow — see Graduated experiments below for the full list and use the Haystack documentation.The remaining experiments were discontinued rather than graduated — see Discontinued experiments. They are not part of Haystack and will not be maintained anywhere.
If you depend on a discontinued experiment, pin the final release explicitly. It will remain installable from PyPI, but it is only tested against the version of Haystack that was current in February 2026 and will drift out of compatibility with newer
haystack-aireleases:pip install "haystack-experimental==0.19.0.post1"
The haystack-experimental package gave Haystack users early access to experimental features without immediately
committing to their official release, so that we could gather feedback and iterate quickly. Each experiment had a
limited lifespan, after which it was either merged into Haystack core, released as a Core Integration, or dropped.
That process has now concluded. This README is kept as the record of where each experiment ended up.
Experiments catalog
Graduated experiments
These experiments were adopted into Haystack core and are available in haystack-ai. The version column is the last
release of haystack-experimental that contained the experimental copy.
| Name | Type | Final release |
|---|---|---|
ChatMessage refactoring; Tool class; tool support in ChatGenerators; ToolInvoker |
Tool Calling support | 0.4.0 |
AsyncPipeline; Pipeline bug fixes and refactoring |
AsyncPipeline execution | 0.7.0 |
LLMMetadataExtractor |
Metadata extraction with LLM | 0.7.0 |
Auto-Merging Retriever & HierarchicalDocumentSplitter |
Document Splitting & Retrieval Technique | 0.8.0 |
Agent |
Simplify Agent development | 0.8.0 |
SuperComponent |
Simplify Pipeline development | 0.8.0 |
Pipeline |
Pipeline breakpoints for debugging | 0.12.0 |
ImageContent; Image Converters; multimodal support in OpenAIChatGenerator and AmazonBedrockChatGenerator; ChatPromptBuilder refactoring; SentenceTransformersDocumentImageEmbedder; LLMDocumentContentExtractor; new Routers |
Multimodality | 0.12.0 |
QueryExpander |
Query Expansion Component | 0.14.3 |
MultiQueryEmbeddingRetriever |
MultiQueryEmbeddingRetriever | 0.14.3 |
MultiQueryTextRetriever |
MultiQueryTextRetriever | 0.14.3 |
EmbeddingBasedDocumentSplitter |
Document Splitting | 0.15.2 |
Confirmation Policies; ConfirmationUIs; BlockingConfirmationStrategy; ConfirmationUIResult; ToolExecutionDecision |
Human in the Loop | 0.16.0 |
Mem0MemoryStore |
MemoryStore | 0.19.0 |
Discontinued experiments
These experiments were not adopted into Haystack. They exist only in the release listed below.
| Name | Type | Final release | Cookbook | Discussion |
|---|---|---|---|---|
OpenAIFunctionCaller |
Function Calling Component | 0.3.0 | None | -- |
OpenAPITool |
OpenAPITool component | 0.3.0 | Notebook | Discuss |
EvaluationHarness |
Evaluation orchestrator | 0.7.0 | None | Discuss |
Agent; BreakpointConfirmationStrategy; HITLBreakpointException |
Human in the Loop via Breakpoints | 0.19.0 | None | Discuss |
InMemoryChatMessageStore; ChatMessageRetriever; ChatMessageWriter |
Chat Message Store, Retriever, Writer | 0.19.0 |  |
Discuss |
OpenAIChatGenerator (hallucination risk scoring) |
Chat Generator Component | 0.19.0 | |
Discuss |
MarkdownHeaderLevelInferrer |
Preprocessor | 0.19.0 | None | Discuss |
LLMSummarizer |
Document Summarizer | 0.19.0 | None | Discuss |
Usage
Experimental features were imported like any other Haystack integration package:
from haystack import Document
from haystack.components.generators.chat import OpenAIChatGenerator
from haystack_experimental.components.summarizers import LLMSummarizer
summarizer = LLMSummarizer(chat_generator=OpenAIChatGenerator())
summarizer.run(documents=[Document(content="...")])
Some experiments came with example notebooks in the Haystack Cookbook.
Telemetry
As with the Haystack core package, this package collected anonymous usage statistics. For more information on what was collected and how to opt out, refer to the telemetry documentation.
1 maintainer
Metadata
1 maintainer
Release files for haystack-experimental 0.19.0.post1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| haystack_experimental-0.19.0.post1.tar.gz | 45.0 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| haystack_experimental-0.19.0.post1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 107.6 kB
Release files / haystack_experimental-0.19.0.post1.tar.gz
| Download URL | haystack_experimental-0.19.0.post1.tar.gz |
|---|---|
| Size | 45.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
898a6974567655b345d0088900bbb6ac539124faec6e209d37307ee7aa6ec652
|
|
BLAKE2b-256 checksum How to use checksums |
a592747d8955f3d6558a16d5b113814a074c05f04e1f6b2a1a1748028bc65032
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 28, 2026.
Transparency logRelease files / haystack_experimental-0.19.0.post1-py3-none-any.whl
| Download URL | haystack_experimental-0.19.0.post1-py3-none-any.whl |
|---|---|
| Size | 62.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
b2d0a9548ff3a16de0138cbfc8dccd94a81eeec1315cc9f3fe304808a138610f
|
|
BLAKE2b-256 checksum How to use checksums |
98b52bc81bbfbfa83ed8a7d123ddc079e1c0ac52a5b6cfd3876bd6ba579cf313
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 28, 2026.
Transparency log
