agent-framework-lib 0.8.9.post3__tar.gz → 0.8.10__tar.gz

agent_framework_lib-0.8.10/LICENSE

@@ -0,0 +1,96 @@
+# PolyForm Internal Use License 1.0.0
+
+<https://polyformproject.org/licenses/internal-use/1.0.0>
+
+## Acceptance
+
+In order to get any license under these terms, you must agree
+to them as both strict obligations and conditions to all
+your licenses.
+
+## Copyright License
+
+The licensor grants you a copyright license for the software
+to do everything you might do with the software that would
+otherwise infringe the licensor's copyright in it for any
+permitted purpose.  However, you may only make changes or
+new works based on the software according to [Changes and New
+Works License](#changes-and-new-works-license), and you may
+not distribute the software.
+
+## Changes and New Works License
+
+The licensor grants you an additional copyright license to
+make changes and new works based on the software for any
+permitted purpose.
+
+## Patent License
+
+The licensor grants you a patent license for the software that
+covers patent claims the licensor can license, or becomes able
+to license, that you would infringe by using the software.
+
+## Fair Use
+
+You may have "fair use" rights for the software under the
+law. These terms do not limit them.
+
+## Internal Business Use
+
+Use of the software for the internal business operations of
+you and your company is use for a permitted purpose.
+
+## No Other Rights
+
+These terms do not allow you to sublicense or transfer any of
+your licenses to anyone else, or prevent the licensor from
+granting licenses to anyone else.  These terms do not imply
+any other licenses.
+
+## Patent Defense
+
+If you make any written claim that the software infringes or
+contributes to infringement of any patent, your patent license
+for the software granted under these terms ends immediately. If
+your company makes such a claim, your patent license ends
+immediately for work on behalf of your company.
+
+## Violations
+
+The first time you are notified in writing that you have
+violated any of these terms, or done anything with the software
+not covered by your licenses, your licenses can nonetheless
+continue if you come into full compliance with these terms,
+and take practical steps to correct past violations, within
+32 days of receiving notice.  Otherwise, all your licenses
+end immediately.
+
+## No Liability
+
+***As far as the law allows, the software comes as is, without
+any warranty or condition, and the licensor will not be liable
+to you for any damages arising out of these terms or the use
+or nature of the software, under any kind of legal claim.***
+
+## Definitions
+
+The **licensor** is the individual or entity offering these
+terms, and the **software** is the software the licensor makes
+available under these terms.
+
+**You** refers to the individual or entity agreeing to these
+terms.
+
+**Your company** is any legal entity, sole proprietorship,
+or other kind of organization that you work for, plus all
+organizations that have control over, are under the control of,
+or are under common control with that organization.  **Control**
+means ownership of substantially all the assets of an entity,
+or the power to direct its management and policies by vote,
+contract, or otherwise.  Control can be direct or indirect.
+
+**Your licenses** are all the licenses granted to you for the
+software under these terms.
+
+**Use** means anything you do with the software requiring one
+of your licenses.

{agent_framework_lib-0.8.9.post3/agent_framework_lib.egg-info → agent_framework_lib-0.8.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-framework-lib
-Version: 0.8.9.post3
+Version: 0.8.10
 Summary: A comprehensive Python framework for building and serving conversational AI agents with FastAPI
 Author-email: Sebastian Pavel <sebastian@cinco.ai>, Elliott Girard <elliott.girard@icloud.com>
 Maintainer-email: Sebastian Pavel <sebastian@cinco.ai>
@@ -60,7 +60,7 @@ Requires-Dist: llama-index>=0.14.16
 Requires-Dist: llama-index-core>=0.14.16
 Requires-Dist: llama-index-workflows>=2.16.0
 Requires-Dist: llama-index-llms-openai>=0.7.9
-Requires-Dist: llama-index-llms-anthropic>=0.11.5
+Requires-Dist: llama-index-llms-anthropic>=0.11.8
 Requires-Dist: llama-index-llms-google-genai>=0.1.0
 Requires-Dist: graphiti-core>=0.24.3
 Requires-Dist: tiktoken>=0.8.0
@@ -83,7 +83,7 @@ Requires-Dist: llama-index-core>=0.14.16; extra == "llamaindex"
 Requires-Dist: llama-index-workflows>=2.16.0; extra == "llamaindex"
 Requires-Dist: llama-index-llms-openai>=0.7.9; extra == "llamaindex"
 Requires-Dist: llama-index-llms-google-genai>=0.1.0; extra == "llamaindex"
-Requires-Dist: llama-index-llms-anthropic>=0.11.5; extra == "llamaindex"
+Requires-Dist: llama-index-llms-anthropic>=0.11.8; extra == "llamaindex"
 Provides-Extra: langgraph
 Requires-Dist: langgraph>=0.2.0; extra == "langgraph"
 Provides-Extra: mcp
@@ -430,7 +430,7 @@ ANTHROPIC_API_KEY=sk-ant-your-anthropic-key
 GEMINI_API_KEY=your-gemini-key
 
 # Model Configuration
-DEFAULT_MODEL=gpt-5.4-mini
+DEFAULT_MODEL=claude-haiku-4-5-20251001
 
 # Multi-Model Routing (Auto Mode)
 DEFAULT_MODEL_MODE=auto                    # "auto" or specific model name
@@ -577,7 +577,7 @@ class MyAgent(LlamaIndexAgent):
         )
         # Default model config (can be overridden per session)
         self.default_temperature = 0.7
-        self.default_model = "gpt-5.4-mini"
+        self.default_model = "claude-haiku-4-5-20251001"
 ```
 
 **Runtime Configuration:**

{agent_framework_lib-0.8.9.post3/agent_framework/web/docs → agent_framework_lib-0.8.10}/README.md

@@ -272,7 +272,7 @@ ANTHROPIC_API_KEY=sk-ant-your-anthropic-key
 GEMINI_API_KEY=your-gemini-key
 
 # Model Configuration
-DEFAULT_MODEL=gpt-5.4-mini
+DEFAULT_MODEL=claude-haiku-4-5-20251001
 
 # Multi-Model Routing (Auto Mode)
 DEFAULT_MODEL_MODE=auto                    # "auto" or specific model name
@@ -419,7 +419,7 @@ class MyAgent(LlamaIndexAgent):
         )
         # Default model config (can be overridden per session)
         self.default_temperature = 0.7
-        self.default_model = "gpt-5.4-mini"
+        self.default_model = "claude-haiku-4-5-20251001"
 ```
 
 **Runtime Configuration:**

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10}/agent_framework/core/base_agent.py

@@ -48,7 +48,7 @@ from .agent_interface import (
     TextOutputPart,
     TextOutputStreamPart,
 )
-from .model_config import model_config
+from .model_config import FRAMEWORK_DEFAULT_MODEL, model_config
 from .models import Tag
 
 
@@ -703,7 +703,7 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
         )
         if not candidate:
             # Final fallback if nothing is configured
-            return "gpt-4o-mini"
+            return FRAMEWORK_DEFAULT_MODEL
         return candidate
 
     def _ensure_agent_built(self):
@@ -1116,7 +1116,7 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
                         user_id,
                         work_profile.total_interactions,
                     )
-                    return "\n\n[Directives de travail]\n" f"{directives}"
+                    return f"\n\n[Directives de travail]\n{directives}"
         except Exception as e:
             logger.debug("Work pattern injection skipped: %s", e)
         return ""
@@ -1241,7 +1241,7 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
         if parts_text:
             for i, text in enumerate(parts_text):
                 logger.info(
-                    f"[_build_full_query] Part {i+1} length: {len(text)} chars, preview: {text[:200]}..."
+                    f"[_build_full_query] Part {i + 1} length: {len(text)} chars, preview: {text[:200]}..."
                 )
 
         # Add timestamp for temporal awareness
@@ -1353,8 +1353,7 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
 
                     if success:
                         logger.info(
-                            "[MEMORY_STORE] Auto-stored interaction for user=%s "
-                            "(response_len=%d)",
+                            "[MEMORY_STORE] Auto-stored interaction for user=%s (response_len=%d)",
                             user_id,
                             len(response_text),
                         )

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10}/agent_framework/core/context_budget.py

@@ -45,9 +45,17 @@ class ModelContextRegistry:
         "gpt-5.4": 1_050_000,
         "gpt-5.5": 400_000,
         "claude-opus-4-6": 200_000,
+        "claude-opus-4-7": 200_000,
         "claude-opus-4-8": 200_000,
         "claude-sonnet-4-6": 1_000_000,
         "claude-opus-4-5-20251101": 200_000,
+        # Claude 5 (disponibles depuis juin-août 2026) : 1M de contexte.
+        "claude-sonnet-5": 1_000_000,
+        "claude-opus-5": 1_000_000,
+        "claude-fable-5": 1_000_000,
+        "claude-fable-5-1": 1_000_000,
+        "gpt-5.4-pro": 1_050_000,
+        "gpt-5.5-pro": 400_000,
         "gemini-2.5-pro": 1_048_576,
         "gemini-3.1-pro-preview": 1_048_576,
     }

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10}/agent_framework/core/model_clients.py

@@ -5,13 +5,20 @@ Creates appropriate model clients (OpenAI, Anthropic, Gemini, etc.) based on mod
 This module is framework-agnostic and uses native client libraries.
 """
 
+import inspect
 import json
 import logging
 import re
 from typing import Any, cast
 
 from .agent_interface import AgentConfig
-from .model_config import ModelConfigManager, ModelProvider, model_config
+from .model_config import (
+    ANTHROPIC_SAMPLING_PARAMS,
+    ModelConfigManager,
+    ModelProvider,
+    anthropic_supports_sampling,
+    model_config,
+)
 
 
 logger = logging.getLogger(__name__)
@@ -175,6 +182,70 @@ def _patch_openai_client_for_sanitization(llm: Any) -> None:
     logger.info("[OpenAI Sanitizer] Patched LlamaIndex OpenAI LLM for cross-provider compatibility")
 
 
+_SAMPLING_DROP_WARNED: set[tuple[str, str]] = set()
+
+
+def _callable_accepts_kwarg(fn: Any, name: str) -> bool | None:
+    """True/False si la signature de ``fn`` tranche, None si on ne peut pas savoir.
+
+    ``None`` (signature introuvable, ou ``**kwargs`` accepté) → on laisse passer :
+    c'est le cas des doubles de test (``AsyncMock``) et des SDK non introspectables.
+    """
+    if fn is None:
+        return None
+    try:
+        params = inspect.signature(fn).parameters
+    except (TypeError, ValueError):
+        return None
+    if any(p.kind is inspect.Parameter.VAR_KEYWORD for p in params.values()):
+        return None
+    return name in params
+
+
+def strip_unsupported_anthropic_kwargs(
+    create_fn: Any, kwargs: dict[str, Any], *, model: str | None = None
+) -> dict[str, Any]:
+    """Retire temperature / top_p / top_k quand le SDK ou le modèle ne les accepte plus.
+
+    Deux sources de vérité, cumulées :
+    1. **le modèle** — Claude Opus 4.7+, Sonnet 5, Opus 5, Fable 5 refusent le sampling
+       (``anthropic_supports_sampling``) ;
+    2. **le SDK** — ``anthropic`` >= 1.0 a retiré ces arguments de ``messages.create`` :
+       la signature réelle fait foi (introspection), ce qui couvre aussi les versions
+       futures sans mise à jour de liste.
+
+    Ne modifie pas ``kwargs`` en place. Journalise une fois par (paramètre, raison).
+    """
+    present = [name for name in ANTHROPIC_SAMPLING_PARAMS if name in kwargs]
+    if not present:
+        return kwargs
+
+    reason: str | None = None
+    if not anthropic_supports_sampling(model):
+        reason = f"modèle {model} sans sampling"
+    else:
+        rejected = [name for name in present if _callable_accepts_kwarg(create_fn, name) is False]
+        if rejected:
+            present = rejected
+            reason = "absent de la signature du SDK anthropic installé"
+
+    if reason is None:
+        return kwargs
+
+    cleaned = {k: v for k, v in kwargs.items() if k not in present}
+    for name in present:
+        key = (name, reason)
+        if key not in _SAMPLING_DROP_WARNED:
+            _SAMPLING_DROP_WARNED.add(key)
+            logger.warning(
+                "[Anthropic] paramètre '%s' ignoré (%s) — valeur %r non envoyée",
+                name,
+                reason,
+                kwargs.get(name),
+            )
+    return cleaned
+
+
 def _patch_anthropic_client_for_sanitization(llm: Any) -> None:
     """
     Patch the LlamaIndex Anthropic LLM to sanitize messages before API calls.
@@ -262,9 +333,18 @@ def _patch_anthropic_client_for_sanitization(llm: Any) -> None:
         original_create = messages_api.create
 
         async def sanitizing_create(*args, **kwargs):
-            """Wrapper that sanitizes messages before calling Anthropic API."""
+            """Wrapper that sanitizes messages before calling Anthropic API.
+
+            Also drops the sampling parameters (temperature / top_p / top_k) that the
+            installed SDK or the target model no longer accepts — llama-index still
+            sends ``temperature`` for pre-Claude-5 models, and ``anthropic`` >= 1.0
+            removed the argument from ``messages.create`` altogether.
+            """
             if "messages" in kwargs:
                 kwargs["messages"] = _sanitize_anthropic_messages(kwargs["messages"])
+            kwargs = strip_unsupported_anthropic_kwargs(
+                original_create, kwargs, model=kwargs.get("model")
+            )
             return await original_create(*args, **kwargs)
 
         messages_api.create = sanitizing_create
@@ -1430,7 +1510,9 @@ def _ensure_openai_json_keyword(
         return messages
     for msg in messages:
         content = msg.get("content")
-        text = content if isinstance(content, str) else (str(content) if content is not None else "")
+        text = (
+            content if isinstance(content, str) else (str(content) if content is not None else "")
+        )
         if "json" in text.lower():
             return messages
     return list(messages) + [
@@ -1634,6 +1716,10 @@ async def _execute_anthropic(
         create_kwargs["system"] = system_text
     if output_config is not None:
         create_kwargs["output_config"] = output_config
+    # SDK anthropic >= 1.0 et modèles Claude >= Opus 4.7 : plus de temperature/top_p/top_k.
+    create_kwargs = strip_unsupported_anthropic_kwargs(
+        getattr(getattr(client, "messages", None), "create", None), create_kwargs, model=model
+    )
 
     if stream:
 

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10}/agent_framework/core/model_config.py

@@ -14,7 +14,7 @@ This module provides:
 Environment Variables:
 - OPENAI_API_KEY: OpenAI API key
 - GEMINI_API_KEY: Google Gemini API key
-- DEFAULT_MODEL: Default model to use (default: "gpt-5.4-mini")
+- DEFAULT_MODEL: Default model to use (default: "claude-haiku-4-5-20251001")
 - OPENAI_MODELS: Comma-separated list of OpenAI model names
 - GEMINI_MODELS: Comma-separated list of Gemini model names
 - FALLBACK_PROVIDER: Provider to use when model provider is unknown
@@ -50,6 +50,12 @@ except ImportError:
 logger = logging.getLogger(__name__)
 
 
+# Modèle par défaut du framework quand rien n'est configuré (DEFAULT_MODEL absent,
+# classifieur du routeur, titres de session, compresseur, agent d'aide…).
+# Décision Komus 15/09/2026 : Claude Haiku 4.5 partout (rapide, peu cher, vision).
+FRAMEWORK_DEFAULT_MODEL: Final[str] = "claude-haiku-4-5-20251001"
+
+
 class ModelProvider(Enum):
     """
     Supported model providers.
@@ -86,11 +92,18 @@ class ModelConfigManager:
     """
 
     # Default model mappings (can be overridden by environment variables)
+    # Modèles constatés sur les comptes Komus le 15/09/2026 (GET /v1/models des deux
+    # fournisseurs). Un id absent d'ici reste routé par motif (« gpt », « claude »).
     DEFAULT_OPENAI_MODELS: Final[list[str]] = [
         "gpt-5.4-nano",
         "gpt-5.4-mini",
         "gpt-5.4",
+        "gpt-5.4-pro",
         "gpt-5.5",
+        "gpt-5.5-pro",
+        "gpt-5.6-sol",
+        "gpt-5.6-terra",
+        "gpt-5.6-luna",
     ]
 
     DEFAULT_ANTHROPIC_MODELS: Final[list[str]] = [
@@ -98,7 +111,12 @@ class ModelConfigManager:
         "claude-sonnet-4-6",
         "claude-opus-4-5-20251101",
         "claude-opus-4-6",
+        "claude-opus-4-7",
         "claude-opus-4-8",
+        "claude-sonnet-5",
+        "claude-opus-5",
+        "claude-fable-5",
+        "claude-fable-5-1",
     ]
 
     DEFAULT_GEMINI_MODELS: Final[list[str]] = [
@@ -150,7 +168,7 @@ class ModelConfigManager:
         self.gemini_api_key = os.getenv("GEMINI_API_KEY", "")
 
         # Default model
-        self.default_model = os.getenv("DEFAULT_MODEL", "gpt-5.4-mini")
+        self.default_model = os.getenv("DEFAULT_MODEL", FRAMEWORK_DEFAULT_MODEL)
 
         # Model mappings from environment (with fallbacks to defaults)
         openai_models_str = os.getenv("OPENAI_MODELS", "")
@@ -532,7 +550,7 @@ class ModelConfigManager:
             ```python
             model_config.register_agent_config("my-agent", {
                 "system_prompt": "You are a helpful assistant.",
-                "model_name": "gpt-5.4-minio",
+                "model_name": "claude-haiku-4-5-20251001",
                 "model_config": {"temperature": 0.7, "max_tokens": 50000}
             })
             ```
@@ -876,3 +894,35 @@ class ModelConfigManager:
 
 # Global instance
 model_config = ModelConfigManager()
+
+
+# ---------------------------------------------------------------------------
+# Anthropic : paramètres de sampling (temperature / top_p / top_k)
+# ---------------------------------------------------------------------------
+# Depuis Claude Opus 4.7 et toute la famille Claude 5, l'API Messages refuse
+# ``temperature``/``top_p``/``top_k`` ; le SDK ``anthropic`` >= 1.0 (09/2026) les a
+# retirés de ``messages.create`` pour TOUS les modèles (TypeError « unexpected keyword
+# argument 'temperature' » constatée en prod le 15/09/2026 sur claude-sonnet-4-6).
+# Les deux cas sont traités dans ``model_clients`` : règle par modèle ci-dessous +
+# introspection de la signature du SDK.
+ANTHROPIC_NO_SAMPLING_MODELS: Final[tuple[str, ...]] = (
+    "claude-opus-4-7",
+    "claude-opus-4-8",
+    "claude-sonnet-5",
+    "claude-opus-5",
+    "claude-fable-5",
+)
+
+ANTHROPIC_SAMPLING_PARAMS: Final[tuple[str, ...]] = ("temperature", "top_p", "top_k")
+
+
+def anthropic_supports_sampling(model_name: str | None) -> bool:
+    """True si le modèle Anthropic accepte encore temperature / top_p / top_k.
+
+    Correspondance par sous-chaîne (comme llama-index) : ``claude-fable-5-1`` et
+    ``anthropic.claude-opus-5-…`` (Bedrock) tombent dans la règle. ``None`` → True.
+    """
+    if not model_name:
+        return True
+    lowered = model_name.lower()
+    return not any(marker in lowered for marker in ANTHROPIC_NO_SAMPLING_MODELS)

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10}/agent_framework/core/model_router.py

@@ -138,14 +138,19 @@ class ModelRouter:
         "gemini-3.1-flash-lite-preview",
     ]
     DEFAULT_STANDARD_MODELS: list[str] = [
+        "claude-sonnet-5",
         "claude-sonnet-4-6",
         "gpt-5.4-mini",
         "gemini-2.5-flash",
         "gemini-3-flash-preview",
     ]
     DEFAULT_ADVANCED_MODELS: list[str] = [
+        "claude-opus-5",
+        "claude-fable-5-1",
         "claude-opus-4-8",
         "gpt-5.5",
+        "gpt-5.5-pro",
+        "claude-sonnet-5",
         "claude-sonnet-4-6",
         "claude-opus-4-6",
         "claude-opus-4-5-20251101",
@@ -193,7 +198,9 @@ class ModelRouter:
         )
 
         # Load classifier and default mode
-        self.classifier_model = os.getenv("AUTO_CLASSIFIER_MODEL", "gpt-4o-mini")
+        from agent_framework.core.model_config import FRAMEWORK_DEFAULT_MODEL
+
+        self.classifier_model = os.getenv("AUTO_CLASSIFIER_MODEL", FRAMEWORK_DEFAULT_MODEL)
         self.default_mode = os.getenv("DEFAULT_MODEL_MODE", "auto")
 
         logger.info("[ModelRouter] Configuration loaded:")

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10}/agent_framework/core/scratchpad_compressor.py

@@ -20,6 +20,7 @@ from agent_framework.core.knowledge_state import (
     ToolCallSignature,
     ToolResultEntry,
 )
+from agent_framework.core.model_config import FRAMEWORK_DEFAULT_MODEL
 
 
 if TYPE_CHECKING:
@@ -55,7 +56,7 @@ class ScratchpadCompressor:
         self,
         token_counter: TokenCounter,
         summarizer: ContextSummarizer | None = None,
-        default_model: str = "gpt-4o-mini",
+        default_model: str = FRAMEWORK_DEFAULT_MODEL,
     ) -> None:
         self._token_counter = token_counter
         self._summarizer = summarizer
@@ -326,9 +327,10 @@ class ScratchpadCompressor:
         effective_model = model_name or self._default_model
         if not effective_model:
             logger.warning(
-                "model_name et default_model sont vides, utilisation de 'gpt-4o-mini' en dernier recours"
+                "model_name et default_model sont vides, utilisation de %s en dernier recours",
+                FRAMEWORK_DEFAULT_MODEL,
             )
-            effective_model = "gpt-4o-mini"
+            effective_model = FRAMEWORK_DEFAULT_MODEL
         if not scratchpad:
             return scratchpad
 

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10}/agent_framework/utils/session_title_generator.py

@@ -12,7 +12,7 @@ Key Features:
 
 Environment Variables:
 - AUTO_GENERATE_SESSION_TITLES: Enable/disable auto-generation (default: "true")
-- SESSION_TITLE_MODEL: LLM model for generation (default: "gpt-5.4-nano")
+- SESSION_TITLE_MODEL: LLM model for generation (default: "claude-haiku-4-5-20251001")
 - SESSION_TITLE_MAX_TOKENS: Max tokens for title (default: 20)
 - SESSION_TITLE_TEMPERATURE: LLM temperature (default: 0.3)
 """
@@ -21,7 +21,8 @@ import logging
 import os
 import re
 
-from ..core.model_clients import OPENAI_AVAILABLE, ModelClientFactory
+from ..core.model_clients import ModelClientFactory, execute_chat_completion
+from ..core.model_config import FRAMEWORK_DEFAULT_MODEL, model_config
 from ..session.session_storage import SessionStorageInterface
 
 
@@ -29,7 +30,7 @@ logger = logging.getLogger(__name__)
 
 # Default configuration values
 DEFAULT_AUTO_GENERATE = True
-DEFAULT_MODEL = "gpt-4o-mini"  # More reliable than gpt-5.4-mini for this task
+DEFAULT_MODEL = FRAMEWORK_DEFAULT_MODEL
 DEFAULT_MAX_TOKENS = 30  # Increased for better title generation
 DEFAULT_TEMPERATURE = 0.3
 DEFAULT_LABEL = "Nouvelle session"
@@ -84,7 +85,7 @@ class SessionTitleGenerator:
         Initialize the session title generator.
 
         Args:
-            model_name: LLM model to use for generation. Defaults to env var or gpt-5.4-nano.
+            model_name: LLM model to use for generation. Defaults to env var or the framework model.
             max_tokens: Maximum tokens for generated title. Defaults to env var or 20.
             temperature: LLM temperature. Defaults to env var or 0.3.
             auto_generate_enabled: Whether auto-generation is enabled. Defaults to env var or True.
@@ -311,10 +312,13 @@ class SessionTitleGenerator:
         """
         effective_model = model_name or self.model_name
 
-        # Check if OpenAI is available (primary provider for title generation)
-        if not OPENAI_AVAILABLE:
+        # Sans clé pour le fournisseur du modèle → extraction simple (pas d'appel LLM).
+        provider = model_config.get_provider_for_model(effective_model)
+        if not model_config.get_api_key_for_provider(provider):
             logger.warning(
-                "[SessionTitleGenerator] OpenAI client not available, using simple extraction"
+                "[SessionTitleGenerator] No API key for %s (model %s), using simple extraction",
+                provider.value,
+                effective_model,
             )
             return self.extract_simple_title(user_message)
 
@@ -329,25 +333,34 @@ class SessionTitleGenerator:
             client = self._client_factory.create_client(
                 model_name=effective_model or self.model_name
             )
-
-            # Use max_completion_tokens for newer models (gpt-5-*), max_tokens for older ones
-            completion_params = {
-                "model": effective_model,
-                "messages": [{"role": "user", "content": prompt}],
-            }
-
-            # gpt-5-* models use max_completion_tokens and only support temperature=1
-            if effective_model.startswith("gpt-5"):
-                completion_params["max_completion_tokens"] = self.max_tokens
-                # gpt-5.4-mini only supports temperature=1 (default)
+            messages = [{"role": "user", "content": prompt}]
+
+            if hasattr(client, "chat"):
+                # SDK OpenAI : chemin historique (max_completion_tokens pour gpt-5-*).
+                completion_params: dict = {"model": effective_model, "messages": messages}
+                if effective_model.startswith("gpt-5"):
+                    completion_params["max_completion_tokens"] = self.max_tokens
+                    # gpt-5.x n'accepte que temperature=1 (défaut)
+                else:
+                    completion_params["max_tokens"] = self.max_tokens
+                    completion_params["temperature"] = self.temperature
+                response = await client.chat.completions.create(**completion_params)
+                raw_title = response.choices[0].message.content
             else:
-                completion_params["max_tokens"] = self.max_tokens
-                completion_params["temperature"] = self.temperature
-
-            response = await client.chat.completions.create(**completion_params)
+                # Anthropic / Gemini : couche d'exécution neutre (gère les paramètres
+                # de sampling refusés par le SDK ou le modèle).
+                model_params = dict(getattr(client, "_model_params", None) or {})
+                model_params["model"] = effective_model
+                result = await execute_chat_completion(
+                    client,
+                    model_params,
+                    messages,
+                    max_tokens=self.max_tokens,
+                    temperature=self.temperature,
+                )
+                raw_title = result.content
 
             # Extract and sanitize the title
-            raw_title = response.choices[0].message.content
             logger.debug(f"[SessionTitleGenerator] Raw LLM response: '{raw_title}'")
             title = self.sanitize_title(raw_title)
 

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10/agent_framework/web/docs}/examples/agent_example_multi_skills.py

@@ -211,7 +211,7 @@ def main() -> None:
     print("=" * 60)
     print("🚀 Starting Multi-Skills Agent Server")
     print("=" * 60)
-    print(f"📊 Model: {os.getenv('DEFAULT_MODEL', 'gpt-4o-mini')}")
+    print(f"📊 Model: {os.getenv('DEFAULT_MODEL', 'claude-haiku-4-5-20251001')}")
     print("🎯 Skills: 12 built-in skills (tools auto-loaded)")
     print(f"🌐 Server: http://localhost:{port}")
     print(f"🎨 UI: http://localhost:{port}/ui")

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10/agent_framework/web/docs}/examples/agent_with_file_storage.py

@@ -159,7 +159,7 @@ def main():
     print("=" * 60)
     print("🚀 Starting File Storage Agent Server")
     print("=" * 60)
-    print(f"📊 Model: {os.getenv('DEFAULT_MODEL', 'gpt-4o-mini')}")
+    print(f"📊 Model: {os.getenv('DEFAULT_MODEL', 'claude-haiku-4-5-20251001')}")
     print("🔧 Tools: create_file, list_files, read_file, create_pdf_from_markdown")
     print("💾 Storage: Local filesystem (default)")
     print(f"🌐 Server: http://localhost:{port}")

{agent_framework_lib-0.8.9.post3 → agent_framework_lib-0.8.10/agent_framework/web/docs}/examples/agent_with_mcp.py

@@ -172,7 +172,7 @@ def main():
     print("=" * 60)
     print("🚀 Starting MCP Agent Server")
     print("=" * 60)
-    print(f"📊 Model: {os.getenv('DEFAULT_MODEL', 'gpt-4o-mini')}")
+    print(f"📊 Model: {os.getenv('DEFAULT_MODEL', 'claude-haiku-4-5-20251001')}")
     print("🔧 Built-in Tools: greet")
     print("🔌 MCP Server: filesystem (if configured)")
     print(f"📁 MCP Directory: {os.getenv('MCP_FILESYSTEM_DIR', '~/mcp_workspace')}")