feat: Add Chat Generator supporting OpenAI Responses API (#9808)

* Add working ChatGenerator

* rename

* Improve and add live tests

* Updates

* Update the tests

* Fix errors

* Add release notes

* Add support for openai tools

* Remove openai tools test that times out

* fix tool calls

* Update release notes

* PR comments

* remove edits to chat message

* Add a test

* PR comments

* Send back reasoning to model

* Fix reasoning support

* Add reasoning support

* Fix tests

* Refactor

* Simplify methods

* Fix mypy

* Stream responses, tool calls etc

* Update docstrings

* Fix errors while using in Agent

* Fix call_id and fc_id

* Update tests

* Updates

* Add extra in ToolCall and ToolCallDelta

* Update streaming chunk

* Fix tests and linting

* Update api key resolve

* PR comments

* PR comments

* Updates

* some type fixes and also make sure to use flatten_tools_or_toolsets

* fix docs

* Fix streaming chunks so assistant header is properly captured

* Add finish_reason and update test

* Skip streaming + pydantic model test b/c of known issue in openai python sdk openai/openai-python#2305

* Fix pylint

---------

Co-authored-by: Sebastian Husch Lee <[email protected]>

Author: Amnah199

docs/pydoc/config/generators_api.yml

@@ -12,6 +12,7 @@ loaders:
         "chat/hugging_face_local",
         "chat/hugging_face_api",
         "chat/openai",
+        "chat/openai_responses",
         "chat/fallback",
       ]
     ignore_when_discovered: ["__init__"]

docs/pydoc/config_docusaurus/generators_api.yml

@@ -12,6 +12,7 @@ loaders:
         "chat/hugging_face_local",
         "chat/hugging_face_api",
         "chat/openai",
+        "chat/openai_responses",
         "chat/fallback",
       ]
     ignore_when_discovered: ["__init__"]

haystack/components/generators/chat/__init__.py

@@ -9,6 +9,7 @@
 
 _import_structure = {
     "openai": ["OpenAIChatGenerator"],
+    "openai_responses": ["OpenAIResponsesChatGenerator"],
     "azure": ["AzureOpenAIChatGenerator"],
     "hugging_face_local": ["HuggingFaceLocalChatGenerator"],
     "hugging_face_api": ["HuggingFaceAPIChatGenerator"],
@@ -21,6 +22,7 @@
     from .hugging_face_api import HuggingFaceAPIChatGenerator as HuggingFaceAPIChatGenerator
     from .hugging_face_local import HuggingFaceLocalChatGenerator as HuggingFaceLocalChatGenerator
     from .openai import OpenAIChatGenerator as OpenAIChatGenerator
+    from .openai_responses import OpenAIResponsesChatGenerator as OpenAIResponsesChatGenerator
 
 else:
     sys.modules[__name__] = LazyImporter(name=__name__, module_file=__file__, import_structure=_import_structure)

haystack/components/generators/chat/openai_responses.py

@@ -30,7 +30,7 @@ class ToolCallDelta:
     index: int
     tool_name: Optional[str] = field(default=None)
     arguments: Optional[str] = field(default=None)
-    id: Optional[str] = field(default=None)  # noqa: A003
+    id: Optional[str] = field(default=None)
     extra: Optional[dict[str, Any]] = field(default=None)
 
     def to_dict(self) -> dict[str, Any]:

haystack/dataclasses/streaming_chunk.py

@@ -0,0 +1,49 @@
+---
+features:
+  - |
+    Added the OpenAIResponsesChatGenerator, a new component that integrates OpenAI's Responses API into Haystack.
+    This unlocks several advanced capabilities from the Responses API:
+    - Allowing retrieval of concise summaries of the model's reasoning process.
+    - Allowing the use of native OpenAI or MCP tool formats, along with Haystack Tool objects and Toolset instances.
+
+    Example with reasoning and web search tool:
+    ```python
+    from haystack.components.generators.chat import OpenAIResponsesChatGenerator
+    from haystack.dataclasses import ChatMessage
+
+    chat_generator = OpenAIResponsesChatGenerator(
+        model="o3-mini",
+        generation_kwargs={
+            {"summary": "auto", "effort": "low"}
+        },
+        tools=[{"type": "web_search"}]
+    )
+
+    response = chat_generator.run(
+        messages=[
+            ChatMessage.from_user("What's a positive news story from today?")
+        ]
+    )
+    print(response["replies"][0].text)
+    ```
+
+    Example with structured output:
+    ```python
+    from pydantic import BaseModel
+    from haystack.components.generators.chat import OpenAIResponsesChatGenerator
+    from haystack.dataclasses import ChatMessage
+
+    class WeatherInfo(BaseModel):
+        location: str
+        temperature: float
+        conditions: str
+
+    chat_generator = OpenAIResponsesChatGenerator(
+        model="gpt-5-mini",
+        generation_kwargs={"text_format": WeatherInfo}
+    )
+
+    response = chat_generator.run(
+        messages=[ChatMessage.from_user("What's the weather in Paris?")]
+    )
+    ```