Skip to content

Use Claude Code, Codex, VSCode, Pi, and OpenCode (and 6 other harnesses) for free (1.3B+ free tokens) from your terminal, app, IDE, or phone, and now from the browser with native browser sessions (multi-harness + multi-model) like OpenClaw (voice supported + ToS friendly)

Python

Provider inferred from model name using prefix

Alishahryar1d6a0e1a401ab
diff --git a/.env.example b/.env.example
index cb9642e9..494daa96 100644
--- a/.env.example
+++ b/.env.example
@@ -1,21 +1,3 @@
-# Provider: "nvidia_nim" | "open_router" | "lmstudio"
-PROVIDER_TYPE="nvidia_nim"
-PROVIDER_RATE_LIMIT=40
-PROVIDER_RATE_WINDOW=60
-# Maximum simultaneous open provider streams (default: 5)
-PROVIDER_MAX_CONCURRENCY=5
-
-
-# HTTP client timeouts (seconds) for provider API requests
-HTTP_READ_TIMEOUT=300
-HTTP_WRITE_TIMEOUT=10
-HTTP_CONNECT_TIMEOUT=2
-
-
-# All Claude model requests are mapped to this model
-MODEL="stepfun-ai/step-3.5-flash"
-
-
# NVIDIA NIM Config
NVIDIA_NIM_API_KEY=""
@@ -28,6 +10,24 @@ OPENROUTER_API_KEY=""
LM_STUDIO_BASE_URL="http://localhost:1234/v1"
+# All Claude model requests are mapped to this model
+# Format: provider_type/model/name
+# Valid providers: "nvidia_nim" | "open_router" | "lmstudio"
+MODEL="nvidia_nim/stepfun-ai/step-3.5-flash"
+
+
+# Provider Config
+PROVIDER_RATE_LIMIT=40
+PROVIDER_RATE_WINDOW=60
+PROVIDER_MAX_CONCURRENCY=5
+
+
+# HTTP client timeouts (seconds) for provider API requests
+HTTP_READ_TIMEOUT=300
+HTTP_WRITE_TIMEOUT=10
+HTTP_CONNECT_TIMEOUT=2
+
+
# Messaging Platform: "telegram" | "discord"
MESSAGING_PLATFORM=discord
MESSAGING_RATE_LIMIT=1
diff --git a/api/models/anthropic.py b/api/models/anthropic.py
index 6c5ebaa6..39da1c55 100644
--- a/api/models/anthropic.py
+++ b/api/models/anthropic.py
@@ -113,7 +113,7 @@ class MessagesRequest(BaseModel):
self.original_model = self.model
# Use centralized model normalization
- normalized = normalize_model_name(self.model, settings.model)
+ normalized = normalize_model_name(self.model, settings.model_name)
if normalized != self.model:
self.model = normalized
@@ -137,4 +137,4 @@ class TokenCountRequest(BaseModel):
"""Map any Claude model name to the configured model."""
settings = get_settings()
# Use centralized model normalization
- return normalize_model_name(v, settings.model)
+ return normalize_model_name(v, settings.model_name)
diff --git a/config/settings.py b/config/settings.py
index 4f367bd7..d77d2081 100644
--- a/config/settings.py
+++ b/config/settings.py
@@ -17,10 +17,6 @@ NVIDIA_NIM_BASE_URL = "https://integrate.api.nvidia.com/v1"
class Settings(BaseSettings):
"""Application settings loaded from environment variables."""
- # ==================== Provider Selection ====================
- # Valid: "nvidia_nim" | "open_router" | "lmstudio"
- provider_type: str = "nvidia_nim"
-
# ==================== OpenRouter Config ====================
open_router_api_key: str = Field(default="", validation_alias="OPENROUTER_API_KEY")
@@ -41,7 +37,8 @@ class Settings(BaseSettings):
# ==================== Model ====================
# All Claude model requests are mapped to this single model
- model: str = "moonshotai/kimi-k2-thinking"
+ # Format: provider_type/model/name
+ model: str = "nvidia_nim/meta/llama3-70b-instruct"
# ==================== Provider Rate Limiting ====================
provider_rate_limit: int = Field(default=40, validation_alias="PROVIDER_RATE_LIMIT")
@@ -124,6 +121,34 @@ class Settings(BaseSettings):
raise ValueError(f"whisper_device must be 'cpu' or 'cuda', got {v!r}")
return v
+ @field_validator("model")
+ @classmethod
+ def validate_model_format(cls, v: str) -> str:
+ valid_providers = ("nvidia_nim", "open_router", "lmstudio")
+ if "/" not in v:
+ raise ValueError(
+ f"Model must be prefixed with provider type. "
+ f"Valid providers: {', '.join(valid_providers)}. "
+ f"Format: provider_type/model/name"
+ )
+ provider = v.split("/", 1)[0]
+ if provider not in valid_providers:
+ raise ValueError(
+ f"Invalid provider: '{provider}'. "
+ f"Supported: 'nvidia_nim', 'open_router', 'lmstudio'"
+ )
+ return v
+
+ @property
+ def provider_type(self) -> str:
+ """Extract provider type from the model string."""
+ return self.model.split("/", 1)[0]
+
+ @property
+ def model_name(self) -> str:
+ """Extract the actual model name from the model string."""
+ return self.model.split("/", 1)[1]
+
model_config = SettingsConfigDict(
env_file=".env",
env_file_encoding="utf-8",
diff --git a/tests/api/test_dependencies.py b/tests/api/test_dependencies.py
index e50dffad..270ce012 100644
--- a/tests/api/test_dependencies.py
+++ b/tests/api/test_dependencies.py
@@ -13,6 +13,7 @@ from providers.open_router import OpenRouterProvider
def _make_mock_settings(**overrides):
"""Create a mock settings object with all required fields for get_provider()."""
mock = MagicMock()
+ mock.model = "nvidia_nim/meta/llama3"
mock.provider_type = "nvidia_nim"
mock.nvidia_nim_api_key = "test_key"
mock.provider_rate_limit = 40
diff --git a/tests/api/test_models_validators.py b/tests/api/test_models_validators.py
index d4fa9ca5..7727f385 100644
--- a/tests/api/test_models_validators.py
+++ b/tests/api/test_models_validators.py
@@ -9,7 +9,7 @@ from config.settings import Settings
@pytest.fixture
def mock_settings():
settings = Settings()
- settings.model = "target-model-from-settings"
+ settings.model = "nvidia_nim/target-model-from-settings"
return settings
diff --git a/tests/conftest.py b/tests/conftest.py
index 291e894f..196825c5 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -8,7 +8,7 @@ import pytest
# Set mock environment BEFORE any imports that use Settings
os.environ.setdefault("NVIDIA_NIM_API_KEY", "test_key")
-os.environ.setdefault("MODEL", "test-model")
+os.environ.setdefault("MODEL", "nvidia_nim/test-model")
os.environ["PTB_TIMEDELTA"] = "1"
# Add project root to path