release: 0.1.0-alpha.13 (#19)

* feat(api): manual updates

Add GRADIENTAI_AGENT_KEY

* feat(api): share chat completion chunk model between chat and agent.chat

* codegen metadata

* feat(api): manual updates

* chore(internal): codegen related update

* chore(internal): bump pinned h11 dep

* chore(package): mark python 3.13 as supported

* fix(parsing): correctly handle nested discriminated unions

* codegen metadata

* chore(readme): fix version rendering on pypi

* fix(client): don't send Content-Type header on GET requests

* feat: clean up environment call outs

* release: 0.1.0-alpha.13

---------

Co-authored-by: stainless-app[bot] <142633134+stainless-app[bot]@users.noreply.github.com>

Author: stainless-app[bot]

.release-please-manifest.json

@@ -1,3 +1,3 @@
 {
-  ".": "0.1.0-alpha.12"
+  ".": "0.1.0-alpha.13"
 }

.stats.yml

@@ -1,4 +1,4 @@
-configured_endpoints: 76
-openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/digitalocean%2Fgradientai-e8b3cbc80e18e4f7f277010349f25e1319156704f359911dc464cc21a0d077a6.yml
+configured_endpoints: 77
+openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/digitalocean%2Fgradientai-391afaae764eb758523b67805cb47ae3bc319dc119d83414afdd66f123ceaf5c.yml
 openapi_spec_hash: c773d792724f5647ae25a5ae4ccec208
-config_hash: 1c936b3bd798c3fcb25479b19efa999a
+config_hash: 0bd094d86a010f7cbd5eb22ef548a29f

CHANGELOG.md

@@ -1,5 +1,30 @@
 # Changelog
 
+## 0.1.0-alpha.13 (2025-07-15)
+
+Full Changelog: [v0.1.0-alpha.12...v0.1.0-alpha.13](https://github.com/digitalocean/gradientai-python/compare/v0.1.0-alpha.12...v0.1.0-alpha.13)
+
+### Features
+
+* **api:** manual updates ([bd6fecc](https://github.com/digitalocean/gradientai-python/commit/bd6feccf97fa5877085783419f11dad04c57d700))
+* **api:** manual updates ([c2b96ce](https://github.com/digitalocean/gradientai-python/commit/c2b96ce3d95cc9b74bffd8d6a499927eefd23b14))
+* **api:** share chat completion chunk model between chat and agent.chat ([d67371f](https://github.com/digitalocean/gradientai-python/commit/d67371f9f4d0761ea03097820bc3e77654b4d2bf))
+* clean up environment call outs ([64ee5b4](https://github.com/digitalocean/gradientai-python/commit/64ee5b449c0195288d0a1dc55d2725e8cdd6afcf))
+
+
+### Bug Fixes
+
+* **client:** don't send Content-Type header on GET requests ([507a342](https://github.com/digitalocean/gradientai-python/commit/507a342fbcc7c801ba36708e56ea2d2a28a1a392))
+* **parsing:** correctly handle nested discriminated unions ([569e473](https://github.com/digitalocean/gradientai-python/commit/569e473d422928597ccf762133d5e52ac9a8665a))
+
+
+### Chores
+
+* **internal:** bump pinned h11 dep ([6f4e960](https://github.com/digitalocean/gradientai-python/commit/6f4e960b6cb838cbf5e50301375fcb4b60a2cfb3))
+* **internal:** codegen related update ([1df657d](https://github.com/digitalocean/gradientai-python/commit/1df657d9b384cb85d27fe839c0dab212a7773f8f))
+* **package:** mark python 3.13 as supported ([1a899b6](https://github.com/digitalocean/gradientai-python/commit/1a899b66a484986672a380e405f09b1ae94b6310))
+* **readme:** fix version rendering on pypi ([6fbe83b](https://github.com/digitalocean/gradientai-python/commit/6fbe83b11a9e3dbb40cf7f9f627abbbd086ee24a))
+
 ## 0.1.0-alpha.12 (2025-07-02)
 
 Full Changelog: [v0.1.0-alpha.11...v0.1.0-alpha.12](https://github.com/digitalocean/gradientai-python/compare/v0.1.0-alpha.11...v0.1.0-alpha.12)

README.md

@@ -1,6 +1,7 @@
 # GradientAI Python API library
 
-[![PyPI version](<https://img.shields.io/pypi/v/c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python.svg?label=pypi%20(stable)>)](https://pypi.org/project/c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python/)
+<!-- prettier-ignore -->
+[![PyPI version](https://img.shields.io/pypi/v/c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python.svg?label=pypi%20(stable))](https://pypi.org/project/c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python/)
 
 The GradientAI Python library provides convenient access to the GradientAI REST API from any Python 3.8+
 application. The library includes type definitions for all request params and response fields,
@@ -73,7 +74,7 @@ client = AsyncGradientAI(
 
 
 async def main() -> None:
-    completion = await client.agents.chat.completions.create(
+    completion = await client.chat.completions.create(
         messages=[
             {
                 "role": "user",
@@ -104,18 +105,17 @@ pip install --pre c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python[aiohttp]
 Then you can enable it by instantiating the client with `http_client=DefaultAioHttpClient()`:
 
 ```python
-import os
 import asyncio
 from gradientai import DefaultAioHttpClient
 from gradientai import AsyncGradientAI
 
 
 async def main() -> None:
     async with AsyncGradientAI(
-        api_key=os.environ.get("GRADIENTAI_API_KEY"),  # This is the default and can be omitted
+        api_key="My API Key",
         http_client=DefaultAioHttpClient(),
     ) as client:
-        completion = await client.agents.chat.completions.create(
+        completion = await client.chat.completions.create(
             messages=[
                 {
                     "role": "user",
@@ -139,7 +139,7 @@ from gradientai import GradientAI
 
 client = GradientAI()
 
-stream = client.agents.chat.completions.create(
+stream = client.chat.completions.create(
     messages=[
         {
             "role": "user",
@@ -160,7 +160,7 @@ from gradientai import AsyncGradientAI
 
 client = AsyncGradientAI()
 
-stream = await client.agents.chat.completions.create(
+stream = await client.chat.completions.create(
     messages=[
         {
             "role": "user",
@@ -192,7 +192,7 @@ from gradientai import GradientAI
 
 client = GradientAI()
 
-completion = client.agents.chat.completions.create(
+completion = client.chat.completions.create(
     messages=[
         {
             "content": "string",
@@ -221,7 +221,7 @@ from gradientai import GradientAI
 client = GradientAI()
 
 try:
-    client.agents.chat.completions.create(
+    client.chat.completions.create(
         messages=[
             {
                 "role": "user",
@@ -272,7 +272,7 @@ client = GradientAI(
 )
 
 # Or, configure per-request:
-client.with_options(max_retries=5).agents.chat.completions.create(
+client.with_options(max_retries=5).chat.completions.create(
     messages=[
         {
             "role": "user",
@@ -303,7 +303,7 @@ client = GradientAI(
 )
 
 # Override per-request:
-client.with_options(timeout=5.0).agents.chat.completions.create(
+client.with_options(timeout=5.0).chat.completions.create(
     messages=[
         {
             "role": "user",
@@ -352,7 +352,7 @@ The "raw" Response object can be accessed by prefixing `.with_raw_response.` to
 from gradientai import GradientAI
 
 client = GradientAI()
-response = client.agents.chat.completions.with_raw_response.create(
+response = client.chat.completions.with_raw_response.create(
     messages=[{
         "role": "user",
         "content": "What is the capital of France?",
@@ -361,7 +361,7 @@ response = client.agents.chat.completions.with_raw_response.create(
 )
 print(response.headers.get('X-My-Header'))
 
-completion = response.parse()  # get the object that `agents.chat.completions.create()` would have returned
+completion = response.parse()  # get the object that `chat.completions.create()` would have returned
 print(completion.choices)
 ```
 
@@ -376,7 +376,7 @@ The above interface eagerly reads the full response body when you make the reque
 To stream the response body, use `.with_streaming_response` instead, which requires a context manager and only reads the response body once you call `.read()`, `.text()`, `.json()`, `.iter_bytes()`, `.iter_text()`, `.iter_lines()` or `.parse()`. In the async client, these are async methods.
 
 ```python
-with client.agents.chat.completions.with_streaming_response.create(
+with client.chat.completions.with_streaming_response.create(
     messages=[
         {
             "role": "user",

api.md

@@ -1,7 +1,7 @@
 # Shared Types
 
 ```python
-from gradientai.types import APILinks, APIMeta, ChatCompletionTokenLogprob
+from gradientai.types import APILinks, APIMeta, ChatCompletionChunk, ChatCompletionTokenLogprob
 ```
 
 # Agents
@@ -65,12 +65,12 @@ Methods:
 Types:
 
 ```python
-from gradientai.types.agents.chat import AgentChatCompletionChunk, CompletionCreateResponse
+from gradientai.types.agents.chat import CompletionCreateResponse
 ```
 
 Methods:
 
-- <code title="post /chat/completions">client.agents.chat.completions.<a href="./src/gradientai/resources/agents/chat/completions.py">create</a>(\*\*<a href="src/gradientai/types/agents/chat/completion_create_params.py">params</a>) -> <a href="./src/gradientai/types/agents/chat/completion_create_response.py">CompletionCreateResponse</a></code>
+- <code title="post /chat/completions?agent=true">client.agents.chat.completions.<a href="./src/gradientai/resources/agents/chat/completions.py">create</a>(\*\*<a href="src/gradientai/types/agents/chat/completion_create_params.py">params</a>) -> <a href="./src/gradientai/types/agents/chat/completion_create_response.py">CompletionCreateResponse</a></code>
 
 ## EvaluationMetrics
 
@@ -260,7 +260,7 @@ Methods:
 Types:
 
 ```python
-from gradientai.types.chat import ChatCompletionChunk, CompletionCreateResponse
+from gradientai.types.chat import CompletionCreateResponse
 ```
 
 Methods:

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python"
-version = "0.1.0-alpha.12"
+version = "0.1.0-alpha.13"
 description = "The official Python library for GradientAI"
 dynamic = ["readme"]
 license = "Apache-2.0"
@@ -24,6 +24,7 @@ classifiers = [
   "Programming Language :: Python :: 3.10",
   "Programming Language :: Python :: 3.11",
   "Programming Language :: Python :: 3.12",
+  "Programming Language :: Python :: 3.13",
   "Operating System :: OS Independent",
   "Operating System :: POSIX",
   "Operating System :: MacOS",
@@ -38,7 +39,7 @@ Homepage = "https://github.com/digitalocean/gradientai-python"
 Repository = "https://github.com/digitalocean/gradientai-python"
 
 [project.optional-dependencies]
-aiohttp = ["aiohttp", "httpx_aiohttp>=0.1.6"]
+aiohttp = ["aiohttp", "httpx_aiohttp>=0.1.8"]
 
 [tool.rye]
 managed = true

requirements-dev.lock

@@ -48,15 +48,15 @@ filelock==3.12.4
 frozenlist==1.6.2
     # via aiohttp
     # via aiosignal
-h11==0.14.0
+h11==0.16.0
     # via httpcore
-httpcore==1.0.2
+httpcore==1.0.9
     # via httpx
 httpx==0.28.1
     # via c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python
     # via httpx-aiohttp
     # via respx
-httpx-aiohttp==0.1.6
+httpx-aiohttp==0.1.8
     # via c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python
 idna==3.4
     # via anyio

requirements.lock

@@ -36,14 +36,14 @@ exceptiongroup==1.2.2
 frozenlist==1.6.2
     # via aiohttp
     # via aiosignal
-h11==0.14.0
+h11==0.16.0
     # via httpcore
-httpcore==1.0.2
+httpcore==1.0.9
     # via httpx
 httpx==0.28.1
     # via c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python
     # via httpx-aiohttp
-httpx-aiohttp==0.1.6
+httpx-aiohttp==0.1.8
     # via c63a5cfe-b235-4fbe-8bbb-82a9e02a482a-python
 idna==3.4
     # via anyio

src/gradientai/_base_client.py

@@ -529,6 +529,15 @@ def _build_request(
             # work around https://github.com/encode/httpx/discussions/2880
             kwargs["extensions"] = {"sni_hostname": prepared_url.host.replace("_", "-")}
 
+        is_body_allowed = options.method.lower() != "get"
+
+        if is_body_allowed:
+            kwargs["json"] = json_data if is_given(json_data) else None
+            kwargs["files"] = files
+        else:
+            headers.pop("Content-Type", None)
+            kwargs.pop("data", None)
+
         # TODO: report this error to httpx
         return self._client.build_request(  # pyright: ignore[reportUnknownMemberType]
             headers=headers,
@@ -540,8 +549,6 @@ def _build_request(
             # so that passing a `TypedDict` doesn't cause an error.
             # https://github.com/microsoft/pyright/issues/3526#event-6715453066
             params=self.qs.stringify(cast(Mapping[str, Any], params)) if params else None,
-            json=json_data if is_given(json_data) else None,
-            files=files,
             **kwargs,
         )
 

src/gradientai/_client.py

@@ -56,13 +56,15 @@ class GradientAI(SyncAPIClient):
     # client options
     api_key: str | None
     inference_key: str | None
+    agent_key: str | None
     agent_domain: str | None
 
     def __init__(
         self,
         *,
         api_key: str | None = None,
         inference_key: str | None = None,
+        agent_key: str | None = None,
         agent_domain: str | None = None,
         base_url: str | httpx.URL | None = None,
         timeout: Union[float, Timeout, None, NotGiven] = NOT_GIVEN,
@@ -88,6 +90,7 @@ def __init__(
         This automatically infers the following arguments from their corresponding environment variables if they are not provided:
         - `api_key` from `GRADIENTAI_API_KEY`
         - `inference_key` from `GRADIENTAI_INFERENCE_KEY`
+        - `agent_key` from `GRADIENTAI_AGENT_KEY`
         """
         if api_key is None:
             api_key = os.environ.get("GRADIENTAI_API_KEY")
@@ -97,6 +100,10 @@ def __init__(
             inference_key = os.environ.get("GRADIENTAI_INFERENCE_KEY")
         self.inference_key = inference_key
 
+        if agent_key is None:
+            agent_key = os.environ.get("GRADIENTAI_AGENT_KEY")
+        self.agent_key = agent_key
+
         self.agent_domain = agent_domain
 
         if base_url is None:
@@ -200,6 +207,7 @@ def copy(
         *,
         api_key: str | None = None,
         inference_key: str | None = None,
+        agent_key: str | None = None,
         agent_domain: str | None = None,
         base_url: str | httpx.URL | None = None,
         timeout: float | Timeout | None | NotGiven = NOT_GIVEN,
@@ -236,6 +244,7 @@ def copy(
         client = self.__class__(
             api_key=api_key or self.api_key,
             inference_key=inference_key or self.inference_key,
+            agent_key=agent_key or self.agent_key,
             agent_domain=agent_domain or self.agent_domain,
             base_url=base_url or self.base_url,
             timeout=self.timeout if isinstance(timeout, NotGiven) else timeout,
@@ -290,13 +299,15 @@ class AsyncGradientAI(AsyncAPIClient):
     # client options
     api_key: str | None
     inference_key: str | None
+    agent_key: str | None
     agent_domain: str | None
 
     def __init__(
         self,
         *,
         api_key: str | None = None,
         inference_key: str | None = None,
+        agent_key: str | None = None,
         agent_domain: str | None = None,
         base_url: str | httpx.URL | None = None,
         timeout: Union[float, Timeout, None, NotGiven] = NOT_GIVEN,
@@ -322,6 +333,7 @@ def __init__(
         This automatically infers the following arguments from their corresponding environment variables if they are not provided:
         - `api_key` from `GRADIENTAI_API_KEY`
         - `inference_key` from `GRADIENTAI_INFERENCE_KEY`
+        - `agent_key` from `GRADIENTAI_AGENT_KEY`
         """
         if api_key is None:
             api_key = os.environ.get("GRADIENTAI_API_KEY")
@@ -331,6 +343,10 @@ def __init__(
             inference_key = os.environ.get("GRADIENTAI_INFERENCE_KEY")
         self.inference_key = inference_key
 
+        if agent_key is None:
+            agent_key = os.environ.get("GRADIENTAI_AGENT_KEY")
+        self.agent_key = agent_key
+
         self.agent_domain = agent_domain
 
         if base_url is None:
@@ -434,6 +450,7 @@ def copy(
         *,
         api_key: str | None = None,
         inference_key: str | None = None,
+        agent_key: str | None = None,
         agent_domain: str | None = None,
         base_url: str | httpx.URL | None = None,
         timeout: float | Timeout | None | NotGiven = NOT_GIVEN,
@@ -470,6 +487,7 @@ def copy(
         client = self.__class__(
             api_key=api_key or self.api_key,
             inference_key=inference_key or self.inference_key,
+            agent_key=agent_key or self.agent_key,
             agent_domain=agent_domain or self.agent_domain,
             base_url=base_url or self.base_url,
             timeout=self.timeout if isinstance(timeout, NotGiven) else timeout,