release: 1.8.0

.github/workflows/ci.yml

@@ -1,12 +1,14 @@
 name: CI
 on:
   push:
-    branches-ignore:
-      - 'generated'
-      - 'codegen/**'
-      - 'integrated/**'
-      - 'stl-preview-head/**'
-      - 'stl-preview-base/**'
+    branches:
+      - '**'
+      - '!integrated/**'
+      - '!stl-preview-head/**'
+      - '!stl-preview-base/**'
+      - '!generated'
+      - '!codegen/**'
+      - 'codegen/stl/**'
   pull_request:
     branches-ignore:
       - 'stl-preview-head/**'

.release-please-manifest.json

@@ -1,3 +1,3 @@
 {
-  ".": "1.7.12"
+  ".": "1.8.0"
 }

.stats.yml

@@ -1,4 +1,4 @@
 configured_endpoints: 125
-openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/letta-ai%2Fletta-sdk-1c44d73b9152645e7b44512a238467b88c993a2c0151d5911b7f98d05583790e.yml
-openapi_spec_hash: e0633c52cd8d694e72211de731d1354a
+openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/letta-ai%2Fletta-sdk-4f81cefe8beedee5e99c60f70160a0c3fe5a8b7e61a4e3eaf4ec77ddbda61d9f.yml
+openapi_spec_hash: 5d46cb41efa2153e3628a530c5f1e64d
 config_hash: 2dd2cc848568d7bec38b1cc774a5900c

CHANGELOG.md

@@ -1,5 +1,30 @@
 # Changelog
 
+## 1.8.0 (2026-03-18)
+
+Full Changelog: [v1.7.12...v1.8.0](https://github.com/letta-ai/letta-python/compare/v1.7.12...v1.8.0)
+
+### Features
+
+* add zai coding provider ([192fc35](https://github.com/letta-ai/letta-python/commit/192fc354c262fca4e4562816198006e92da1b1b4))
+* always create otid for request ([f208278](https://github.com/letta-ai/letta-python/commit/f20827887378a6a0525d02731436c820e3c6ee06))
+
+
+### Bug Fixes
+
+* **deps:** bump minimum typing-extensions version ([fcebea1](https://github.com/letta-ai/letta-python/commit/fcebea132a9cfa2f4d87ab298ccdc32cf58f6884))
+* **pydantic:** do not pass `by_alias` unless set ([6fd6fc5](https://github.com/letta-ai/letta-python/commit/6fd6fc5ab183d7a0eef902bb07f3ea4801b1e247))
+
+
+### Chores
+
+* **internal:** tweak CI branches ([0591087](https://github.com/letta-ai/letta-python/commit/05910876a703bbd272924a3c54d9a8fe2c993c94))
+
+
+### Refactors
+
+* **sglang:** use input_ids instead of text in native adapter ([66c6a22](https://github.com/letta-ai/letta-python/commit/66c6a223d1c3aa6b081289401185d528218efdba))
+
 ## 1.7.12 (2026-03-09)
 
 Full Changelog: [v1.7.11...v1.7.12](https://github.com/letta-ai/letta-python/compare/v1.7.11...v1.7.12)

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "letta-client"
-version = "1.7.12"
+version = "1.8.0"
 description = "The official Python library for the letta API"
 dynamic = ["readme"]
 license = "Apache-2.0"
@@ -11,7 +11,7 @@ authors = [
 dependencies = [
   "httpx>=0.23.0, <1",
   "pydantic>=1.9.0, <3",
-  "typing-extensions>=4.10, <5",
+  "typing-extensions>=4.14, <5",
   "anyio>=3.5.0, <5",
   "distro>=1.7.0, <2",
   "sniffio",

src/letta_client/_compat.py

@@ -2,7 +2,7 @@
 
 from typing import TYPE_CHECKING, Any, Union, Generic, TypeVar, Callable, cast, overload
 from datetime import date, datetime
-from typing_extensions import Self, Literal
+from typing_extensions import Self, Literal, TypedDict
 
 import pydantic
 from pydantic.fields import FieldInfo
@@ -131,6 +131,10 @@ def model_json(model: pydantic.BaseModel, *, indent: int | None = None) -> str:
     return model.model_dump_json(indent=indent)
 
 
+class _ModelDumpKwargs(TypedDict, total=False):
+    by_alias: bool
+
+
 def model_dump(
     model: pydantic.BaseModel,
     *,
@@ -142,14 +146,17 @@ def model_dump(
     by_alias: bool | None = None,
 ) -> dict[str, Any]:
     if (not PYDANTIC_V1) or hasattr(model, "model_dump"):
+        kwargs: _ModelDumpKwargs = {}
+        if by_alias is not None:
+            kwargs["by_alias"] = by_alias
         return model.model_dump(
             mode=mode,
             exclude=exclude,
             exclude_unset=exclude_unset,
             exclude_defaults=exclude_defaults,
             # warnings are not supported in Pydantic v1
             warnings=True if PYDANTIC_V1 else warnings,
-            by_alias=by_alias,
+            **kwargs,
         )
     return cast(
         "dict[str, Any]",

src/letta_client/_version.py

@@ -1,4 +1,4 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
 __title__ = "letta_client"
-__version__ = "1.7.12"  # x-release-please-version
+__version__ = "1.8.0"  # x-release-please-version

src/letta_client/types/agent_create_params.py

@@ -38,6 +38,9 @@
     "AgentCreateParams",
     "CompactionSettings",
     "CompactionSettingsModelSettings",
+    "CompactionSettingsModelSettingsSgLangModelSettings",
+    "CompactionSettingsModelSettingsSgLangModelSettingsReasoning",
+    "CompactionSettingsModelSettingsSgLangModelSettingsResponseFormat",
     "CompactionSettingsModelSettingsZaiModelSettings",
     "CompactionSettingsModelSettingsZaiModelSettingsResponseFormat",
     "CompactionSettingsModelSettingsZaiModelSettingsThinking",
@@ -46,6 +49,9 @@
     "CompactionSettingsModelSettingsChatGptoAuthModelSettings",
     "CompactionSettingsModelSettingsChatGptoAuthModelSettingsReasoning",
     "ModelSettings",
+    "ModelSettingsSgLangModelSettings",
+    "ModelSettingsSgLangModelSettingsReasoning",
+    "ModelSettingsSgLangModelSettingsResponseFormat",
     "ModelSettingsZaiModelSettings",
     "ModelSettingsZaiModelSettingsResponseFormat",
     "ModelSettingsZaiModelSettingsThinking",
@@ -264,6 +270,51 @@ class AgentCreateParams(TypedDict, total=False):
     """The tools used by the agent."""
 
 
+class CompactionSettingsModelSettingsSgLangModelSettingsReasoning(TypedDict, total=False):
+    """The reasoning configuration for the model."""
+
+    reasoning_effort: Literal["none", "minimal", "low", "medium", "high", "xhigh"]
+    """The reasoning effort to use when generating text reasoning models"""
+
+
+CompactionSettingsModelSettingsSgLangModelSettingsResponseFormat: TypeAlias = Union[
+    TextResponseFormatParam, JsonSchemaResponseFormatParam, JsonObjectResponseFormatParam
+]
+
+
+class CompactionSettingsModelSettingsSgLangModelSettings(TypedDict, total=False):
+    """
+    SGLang model configuration (OpenAI-compatible runtime with SGLang-specific parsing).
+    """
+
+    max_output_tokens: int
+    """The maximum number of tokens the model can generate."""
+
+    parallel_tool_calls: bool
+    """Whether to enable parallel tool calling."""
+
+    provider_type: Literal["sglang"]
+    """The type of the provider."""
+
+    reasoning: CompactionSettingsModelSettingsSgLangModelSettingsReasoning
+    """The reasoning configuration for the model."""
+
+    response_format: Optional[CompactionSettingsModelSettingsSgLangModelSettingsResponseFormat]
+    """The response format for the model."""
+
+    strict: bool
+    """Enable strict mode for tool calling.
+
+    When true, tool outputs are guaranteed to match JSON schemas.
+    """
+
+    temperature: float
+    """The temperature of the model."""
+
+    tool_call_parser: Optional[str]
+    """SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes')."""
+
+
 CompactionSettingsModelSettingsZaiModelSettingsResponseFormat: TypeAlias = Union[
     TextResponseFormatParam, JsonSchemaResponseFormatParam, JsonObjectResponseFormatParam
 ]
@@ -353,6 +404,7 @@ class CompactionSettingsModelSettingsChatGptoAuthModelSettings(TypedDict, total=
 
 CompactionSettingsModelSettings: TypeAlias = Union[
     OpenAIModelSettingsParam,
+    CompactionSettingsModelSettingsSgLangModelSettings,
     AnthropicModelSettingsParam,
     GoogleAIModelSettingsParam,
     GoogleVertexModelSettingsParam,
@@ -409,6 +461,51 @@ class CompactionSettings(TypedDict, total=False):
     """
 
 
+class ModelSettingsSgLangModelSettingsReasoning(TypedDict, total=False):
+    """The reasoning configuration for the model."""
+
+    reasoning_effort: Literal["none", "minimal", "low", "medium", "high", "xhigh"]
+    """The reasoning effort to use when generating text reasoning models"""
+
+
+ModelSettingsSgLangModelSettingsResponseFormat: TypeAlias = Union[
+    TextResponseFormatParam, JsonSchemaResponseFormatParam, JsonObjectResponseFormatParam
+]
+
+
+class ModelSettingsSgLangModelSettings(TypedDict, total=False):
+    """
+    SGLang model configuration (OpenAI-compatible runtime with SGLang-specific parsing).
+    """
+
+    max_output_tokens: int
+    """The maximum number of tokens the model can generate."""
+
+    parallel_tool_calls: bool
+    """Whether to enable parallel tool calling."""
+
+    provider_type: Literal["sglang"]
+    """The type of the provider."""
+
+    reasoning: ModelSettingsSgLangModelSettingsReasoning
+    """The reasoning configuration for the model."""
+
+    response_format: Optional[ModelSettingsSgLangModelSettingsResponseFormat]
+    """The response format for the model."""
+
+    strict: bool
+    """Enable strict mode for tool calling.
+
+    When true, tool outputs are guaranteed to match JSON schemas.
+    """
+
+    temperature: float
+    """The temperature of the model."""
+
+    tool_call_parser: Optional[str]
+    """SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes')."""
+
+
 ModelSettingsZaiModelSettingsResponseFormat: TypeAlias = Union[
     TextResponseFormatParam, JsonSchemaResponseFormatParam, JsonObjectResponseFormatParam
 ]
@@ -498,6 +595,7 @@ class ModelSettingsChatGptoAuthModelSettings(TypedDict, total=False):
 
 ModelSettings: TypeAlias = Union[
     OpenAIModelSettingsParam,
+    ModelSettingsSgLangModelSettings,
     AnthropicModelSettingsParam,
     GoogleAIModelSettingsParam,
     GoogleVertexModelSettingsParam,