Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/shobcoder/shob/llms.txt

Use this file to discover all available pages before exploring further.

Shob works with multiple AI providers out of the box. The provider block in shob.json lets you supply API keys, override base URLs, set request timeouts, control which models are visible, and even define entirely custom OpenAI-compatible endpoints — all without touching environment variables.

​
The provider config block

Each key inside provider is a provider ID (e.g. anthropic, openai). The value is a provider configuration object: 
{
  "provider": {
    "<provider-id>": {
      "options": {
        "apiKey": "...",
        "baseURL": "..."
      }
    }
  }
}

​
Provider object fields

​
options.apiKey
string
API key sent with every request to this provider. Overrides the provider’s default environment-variable lookup.
​
options.baseURL
string
Override the HTTP base URL for all requests. Required when connecting to a custom or self-hosted OpenAI-compatible endpoint.
​
options.timeout
number | false
Request timeout in milliseconds (default: 300000 — 5 minutes). Set to false to disable the timeout entirely.
​
options.chunkTimeout
number
Maximum milliseconds to wait between streamed SSE chunks before aborting. Useful for detecting stalled streams.
​
options.headerTimeout
number | false
Milliseconds to wait for the initial response headers. Set to false to disable.
​
options.setCacheKey
boolean
Enable promptCacheKey for this provider (default: false). Only relevant for providers that support prompt caching.
​
whitelist
string[]
If set, only models in this list will appear in the model selector for this provider.
​
blacklist
string[]
Models to hide from the model selector for this provider.
​
models
object
Per-model metadata overrides (cost, capabilities, etc.). Advanced use only.

​
Setting the default model

The top-level model field sets the default model used for every new session. Use the provider/model-name format: 
{
  "model": "anthropic/claude-opus-4-5"
}
You can also set a small_model for lightweight background tasks like generating session titles: 
{
  "model": "anthropic/claude-opus-4-5",
  "small_model": "anthropic/claude-haiku-3-5"
}
The model picker in the TUI lets you change models per-session at any time. The model field in shob.json is only the startup default.

​
Provider examples

​
Anthropic

{
  "model": "anthropic/claude-opus-4-5",
  "provider": {
    "anthropic": {
      "options": {
        "apiKey": "sk-ant-api03-..."
      }
    }
  }
}
Shob will also read ANTHROPIC_API_KEY from the environment if apiKey is not set in the config.

​
OpenAI

{
  "model": "openai/gpt-4o",
  "provider": {
    "openai": {
      "options": {
        "apiKey": "sk-..."
      }
    }
  }
}
The OPENAI_API_KEY environment variable is the fallback if apiKey is omitted.

​
Custom / OpenAI-compatible endpoint

Any provider that exposes an OpenAI-compatible API (Ollama, LM Studio, vLLM, Together AI, etc.) can be registered with a custom baseURL: 
{
  "model": "custom-local/llama-3.1-8b",
  "provider": {
    "custom-local": {
      "options": {
        "baseURL": "http://localhost:11434/v1",
        "apiKey": "ollama"
      }
    }
  }
}
When using a custom provider ID, set model to your-provider-id/model-name so Shob knows which provider to route requests to.

​
Restricting available models

Use whitelist to limit the models shown in the picker, or blacklist to hide specific ones: 
{
  "provider": {
    "anthropic": {
      "options": {
        "apiKey": "sk-ant-..."
      },
      "whitelist": [
        "claude-opus-4-5",
        "claude-haiku-3-5"
      ]
    },
    "openai": {
      "options": {
        "apiKey": "sk-..."
      },
      "blacklist": [
        "gpt-3.5-turbo"
      ]
    }
  }
}

​
Disabling providers entirely

Use disabled_providers to stop certain providers from loading at all, or enabled_providers to use an allowlist: 
{
  // Disable a specific provider
  "disabled_providers": ["bedrock"],

  // Or allowlist — only these providers will be loaded
  "enabled_providers": ["anthropic", "openai"]
}

​
API keys via environment variables

You can always supply API keys through environment variables instead of (or in addition to) the config file. The standard variables Shob recognises are the same ones each provider’s SDK uses:
ProviderEnvironment variable
AnthropicANTHROPIC_API_KEY
OpenAIOPENAI_API_KEY
Google (Gemini)GOOGLE_GENERATIVE_AI_API_KEY
Amazon BedrockAWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY
Avoid committing API keys directly in shob.json files that are checked into source control. For project configs, prefer environment variables or keep secrets only in your global ~/.config/shob/shob.json.

​
Full provider config reference

{
  "$schema": "https://shob.ai/config.json",

  "model": "anthropic/claude-opus-4-5",
  "small_model": "anthropic/claude-haiku-3-5",

  "provider": {
    // Anthropic with a custom prompt-cache key
    "anthropic": {
      "options": {
        "apiKey": "sk-ant-...",
        "setCacheKey": true,
        "timeout": 600000
      },
      "whitelist": ["claude-opus-4-5", "claude-sonnet-4-5", "claude-haiku-3-5"]
    },

    // OpenAI with a short header timeout
    "openai": {
      "options": {
        "apiKey": "sk-...",
        "headerTimeout": 10000
      }
    },

    // Local Ollama instance (OpenAI-compatible)
    "ollama": {
      "options": {
        "baseURL": "http://localhost:11434/v1",
        "apiKey": "ollama",
        "timeout": false
      }
    }
  }
}

Build docs developers (and LLMs) love

Get started for freeTalk to us