Skip to main content
ScryCLI supports multiple AI model providers through OpenRouter, offering a range of free models optimized for code generation and understanding.

​
Supported Providers

ScryCLI integrates with the following AI providers via OpenRouter:
OpenAI models available through OpenRouter:
  • gpt-oss-120b (Recommended) - openai/gpt-oss-120b:free
  • gpt-oss-20b - openai/gpt-oss-20b:free

​
Model Selection

On first launch, ScryCLI prompts you to select a provider and model through an interactive menu.

​
Selection Flow

const providers = [
  { label: "OpenAI", value: "openai" },
  { label: "StepFun ", value: "stepfun" },
  { label: "Qwen", value: "qwen" },
  { label: "ArceeAI", value: "arceeai" },
  { label: "Mistral", value: "mistral" },
  { label: "Zai", value: "zai" },
];

const modelsMap: Record<string, any[]> = {
  openai: [
    { label: "gpt-oss-120b (Recommended)", value: "openai/gpt-oss-120b:free" },
    { label: "gpt-oss-20b", value: "openai/gpt-oss-20b:free" },
  ],
  // ... other providers
};
The SelectModel component handles the interactive selection: 
const SelectModel = (props: selectModelType) => {
  const [selectedProvider, setSelectedProvider] = useState<string | null>(null);
  const [selectedModel, setSelectedModel] = useState<string | null>(null);

  return (
    <Box>
      {!selectedProvider && (
        <SelectInput
          items={providers}
          onSelect={(item) => setSelectedProvider(item.value)}
        />
      )}

      {selectedProvider && !selectedModel && (
        <SelectInput
          items={modelsMap[selectedProvider] || []}
          onSelect={(item) => {
            setSelectedModel(item.value as string);
            setConfig('model', { 
              modelProvider: selectedProvider, 
              modelName: item.value 
            });
          }}
        />
      )}
    </Box>
  );
}
Your model selection is saved to ~/.scrycli/config.json and persists across sessions.

​
Model Validation

Before allowing chat interaction, ScryCLI validates that a model is properly configured: 
const isModelSelected = (): boolean => {
  try {
    const config = getConfig();
    if (!config?.model?.modelName) return false;
    if (!config?.model?.modelKey) return false;
    return config.model.modelName !== '' && config.model.modelKey !== '';
  } catch(error) {
    console.error(error);
    return false;
  }
};
The validation checks for:
  • model.modelName - The full model identifier (e.g., openai/gpt-oss-120b:free)
  • model.modelKey - API key for authentication
Both modelName and modelKey must be present and non-empty for the chat interface to activate.

​
OpenRouter Integration

ScryCLI uses the OpenRouter SDK to communicate with all AI models: 
import { OpenRouter } from '@openrouter/sdk';

const openRouterClient = new OpenRouter({
  apiKey: '',
});

export async function llmCall({
  prompt,
  systemPrompt,
}: llmCallParams): Promise<string> {
  const result = openRouterClient.callModel({
    model: `${getConfig().model.modelName}`,
    instructions: `${systemPrompt}`,
    input: `${prompt} \n\nFile Tree: ${fileTreeString}`,
  });
  const text = await result.getText();
  return text;
}

​
Why OpenRouter?

OpenRouter provides:
  • Unified API - Single interface for multiple AI providers
  • Free Tier Models - Access to powerful models at no cost
  • Fallback Support - Automatic routing if a model is unavailable
  • Usage Tracking - Monitor API consumption across providers

​
Configuration Storage

Model configuration is stored in the user’s home directory: 
{
  "model": {
    "modelProvider": "openai",
    "modelName": "openai/gpt-oss-120b:free",
    "modelKey": "your-api-key-here"
  }
}
The config file is located at ~/.scrycli/config.json

​
Current Model Display

The active model is displayed in the chat interface: 
<Box alignSelf="flex-end">
  <Text color='yellow'>
    {`Model: ${config?.model?.modelName ?? 'Not Selected'}`}
  </Text>
</Box>
This provides constant visibility into which AI model is processing your requests.

​
Switching Models

To change your model selection:
  1. Exit ScryCLI
  2. Delete or modify ~/.scrycli/config.json
  3. Restart ScryCLI to see the model selection menu again
Alternatively, you can manually edit the config file: 
# Open config in your editor
vim ~/.scrycli/config.json

# Change the modelName value
{
  "model": {
    "modelProvider": "qwen",
    "modelName": "qwen/qwen3-next-80b-a3b-instruct:free",
    "modelKey": "your-api-key"
  }
}

​
Best Practices

Recommended Models for Code:
  • Qwen3 Next 80B - Best for complex code generation
  • OpenAI GPT-OSS-120B - Balanced performance and speed
  • ArceeAI Trinity Large - Good for code understanding tasks
API Key Security:
  • Never commit your config.json to version control
  • Keep your OpenRouter API key confidential
  • Rotate keys regularly for security