code-puppy 0.0.640__tar.gz → 0.0.641__tar.gz

{code_puppy-0.0.640 → code_puppy-0.0.641}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code-puppy
-Version: 0.0.640
+Version: 0.0.641
 Summary: Code generation agent
 Project-URL: repository, https://github.com/mpfaffenberger/code_puppy
 Project-URL: HomePage, https://github.com/mpfaffenberger/code_puppy
@@ -310,6 +310,23 @@ Then just use /model and tab to select your round-robin model!
 
 The `rotate_every` parameter controls how many requests are made to each model before rotating to the next one. In this example, the round-robin model will use each Qwen model for 5 consecutive requests before moving to the next model in the sequence.
 
+## Custom OpenAI API Types
+
+Use `custom_openai` for OpenAI-compatible Chat Completions endpoints. If an endpoint requires the OpenAI Responses API, use `custom_openai_responses` instead:
+
+```json
+{
+  "reasoning_proxy": {
+    "type": "custom_openai_responses",
+    "name": "gpt-5.5",
+    "custom_endpoint": {
+      "url": "https://proxy.example.com/v1",
+      "api_key": "$API_KEY"
+    }
+  }
+}
+```
+
 ## Custom Model Timeouts
 
 For custom model endpoints (`custom_openai`, `custom_anthropic`, `custom_gemini`, `cerebras`), you can configure custom timeout values to handle slow or unreliable endpoints. The default timeout for these custom endpoint models is 180 seconds.

{code_puppy-0.0.640 → code_puppy-0.0.641}/README.md

@@ -265,6 +265,23 @@ Then just use /model and tab to select your round-robin model!
 
 The `rotate_every` parameter controls how many requests are made to each model before rotating to the next one. In this example, the round-robin model will use each Qwen model for 5 consecutive requests before moving to the next model in the sequence.
 
+## Custom OpenAI API Types
+
+Use `custom_openai` for OpenAI-compatible Chat Completions endpoints. If an endpoint requires the OpenAI Responses API, use `custom_openai_responses` instead:
+
+```json
+{
+  "reasoning_proxy": {
+    "type": "custom_openai_responses",
+    "name": "gpt-5.5",
+    "custom_endpoint": {
+      "url": "https://proxy.example.com/v1",
+      "api_key": "$API_KEY"
+    }
+  }
+}
+```
+
 ## Custom Model Timeouts
 
 For custom model endpoints (`custom_openai`, `custom_anthropic`, `custom_gemini`, `cerebras`), you can configure custom timeout values to handle slow or unreliable endpoints. The default timeout for these custom endpoint models is 180 seconds.

{code_puppy-0.0.640 → code_puppy-0.0.641}/code_puppy/model_factory.py

@@ -59,6 +59,18 @@ _load_plugin_model_providers()
 
 # Anthropic beta header required for 1M context window support.
 CONTEXT_1M_BETA = "context-1m-2025-08-07"
+_CUSTOM_OPENAI_MODEL_TYPES = {"custom_openai", "custom_openai_responses"}
+_LEGACY_CUSTOM_OPENAI_RESPONSES_MODEL = "codex-gpt-5-codex"
+
+
+def _custom_openai_uses_responses_api(
+    model_name: str, model_config: Dict[str, Any]
+) -> bool:
+    """Return whether a custom OpenAI model should use the Responses API."""
+    return (
+        model_config.get("type") == "custom_openai_responses"
+        or model_name == _LEGACY_CUSTOM_OPENAI_RESPONSES_MODEL
+    )
 
 
 def _build_anthropic_beta_header(
@@ -287,7 +299,10 @@ def make_model_settings(
             model_type == "chatgpt_oauth"
             or model_type == "azure_foundry_openai"
             or (model_type == "openai" and "codex" in model_name)
-            or (model_type == "custom_openai" and "codex" in model_name)
+            or (
+                model_type in _CUSTOM_OPENAI_MODEL_TYPES
+                and _custom_openai_uses_responses_api(model_name, model_config)
+            )
         )
 
         if uses_responses_api:
@@ -796,7 +811,7 @@ class ModelFactory:
             )
             return OpenAIChatModel(model_name=model_config["name"], provider=provider)
 
-        elif model_type == "custom_openai":
+        elif model_type in _CUSTOM_OPENAI_MODEL_TYPES:
             url, headers, verify, api_key, timeout = get_custom_config(model_config)
             client = create_async_client(
                 headers=headers,
@@ -809,14 +824,15 @@ class ModelFactory:
             if api_key:
                 provider_args["api_key"] = api_key
             provider = make_openai_provider(provider_identity, **provider_args)
-            model = OpenAIChatModel(
+            if _custom_openai_uses_responses_api(model_name, model_config):
+                return OpenAIResponsesModel(
+                    model_name=model_config["name"], provider=provider
+                )
+            return OpenAIChatModel(
                 model_name=model_config["name"],
                 provider=provider,
                 profile=_thinking_tags_profile(model_name, model_config),
             )
-            if model_name == "codex-gpt-5-codex":
-                model = OpenAIResponsesModel(model_config["name"], provider=provider)
-            return model
         elif model_type == "zai_coding":
             api_key = get_api_key("ZAI_API_KEY")
             if not api_key:

{code_puppy-0.0.640 → code_puppy-0.0.641}/code_puppy/provider_identity.py

@@ -88,8 +88,8 @@ def resolve_provider_identity(model_key: str, model_config: dict[str, Any]) -> s
 
     if model_type == "custom_anthropic":
         return "custom_anthropic"
-    if model_type == "custom_openai":
-        return "custom_openai"
+    if model_type in {"custom_openai", "custom_openai_responses"}:
+        return model_type
 
     return model_type or "unknown"
 

{code_puppy-0.0.640 → code_puppy-0.0.641}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "code-puppy"
-version = "0.0.640"
+version = "0.0.641"
 description = "Code generation agent"
 readme = "README.md"
 requires-python = ">=3.11,<3.15"