> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/usestrix/strix/llms.txt
> Use this file to discover all available pages before exploring further.

# Environment Variables

> Configure Strix using environment variables for LLM providers, API access, and runtime behavior

You can configure Strix using environment variables. Strix automatically saves your configuration to `~/.strix/cli-config.json`, so you don't need to set these variables on every run.

## Required Variables

<ParamField path="STRIX_LLM" type="string" required>
  Model name to use with LiteLLM (e.g., `openai/gpt-5`, `anthropic/claude-sonnet-4-6`).

  For Strix Router, use the `strix/` prefix (e.g., `strix/gpt-5`).

  ```bash theme={null}
  export STRIX_LLM="openai/gpt-5"
  ```
</ParamField>

## LLM Configuration

<ParamField path="LLM_API_KEY" type="string">
  API key for your LLM provider.

  Not required for local models (Ollama, LMStudio), cloud providers with built-in authentication (Vertex AI, AWS Bedrock), or when using Strix Router.

  ```bash theme={null}
  export LLM_API_KEY="your-api-key-here"
  ```
</ParamField>

<ParamField path="LLM_API_BASE" type="string">
  Custom API base URL for your LLM provider.

  Use this when running local models (Ollama, LMStudio) or custom endpoints. For `strix/` models, this is automatically set to Strix Router.

  ```bash theme={null}
  # For Ollama
  export LLM_API_BASE="http://localhost:11434"

  # For LMStudio
  export LLM_API_BASE="http://localhost:1234/v1"
  ```
</ParamField>

<ParamField path="OPENAI_API_BASE" type="string">
  OpenAI-specific API base URL. This is an alternative to `LLM_API_BASE` for OpenAI-compatible endpoints.

  ```bash theme={null}
  export OPENAI_API_BASE="https://api.openai.com/v1"
  ```
</ParamField>

<ParamField path="LITELLM_BASE_URL" type="string">
  LiteLLM-specific base URL. This is an alternative to `LLM_API_BASE` when using LiteLLM proxy.

  ```bash theme={null}
  export LITELLM_BASE_URL="http://localhost:4000"
  ```
</ParamField>

<ParamField path="OLLAMA_API_BASE" type="string">
  Ollama-specific API base URL. This is an alternative to `LLM_API_BASE` when using Ollama.

  ```bash theme={null}
  export OLLAMA_API_BASE="http://localhost:11434"
  ```
</ParamField>

<ParamField path="STRIX_REASONING_EFFORT" type="string" default="high">
  Control the reasoning effort level for LLM responses.

  **Valid values:** `none`, `minimal`, `low`, `medium`, `high`, `xhigh`

  Higher values result in more thorough analysis but slower responses. For quick scans, use `medium`.

  ```bash theme={null}
  export STRIX_REASONING_EFFORT="high"
  ```
</ParamField>

<ParamField path="STRIX_LLM_MAX_RETRIES" type="string" default="5">
  Maximum number of retries for LLM API calls when encountering transient errors.

  ```bash theme={null}
  export STRIX_LLM_MAX_RETRIES="5"
  ```
</ParamField>

<ParamField path="LLM_TIMEOUT" type="string" default="300">
  Timeout in seconds for LLM API requests.

  ```bash theme={null}
  export LLM_TIMEOUT="300"
  ```
</ParamField>

<ParamField path="STRIX_MEMORY_COMPRESSOR_TIMEOUT" type="string" default="30">
  Timeout in seconds for the memory compression operation.

  ```bash theme={null}
  export STRIX_MEMORY_COMPRESSOR_TIMEOUT="30"
  ```
</ParamField>

## Tool & Feature Configuration

<ParamField path="PERPLEXITY_API_KEY" type="string">
  API key for Perplexity AI web search.

  Enables real-time research and OSINT capabilities during penetration testing. Get your API key at [perplexity.ai](https://www.perplexity.ai/).

  ```bash theme={null}
  export PERPLEXITY_API_KEY="your-perplexity-key-here"
  ```
</ParamField>

<ParamField path="STRIX_DISABLE_BROWSER" type="string" default="false">
  Disable browser automation tool.

  Set to `true` to prevent Strix from using browser automation (Playwright). This disables testing for XSS, CSRF, and other client-side vulnerabilities.

  ```bash theme={null}
  export STRIX_DISABLE_BROWSER="true"
  ```
</ParamField>

## Runtime Configuration

<ParamField path="STRIX_IMAGE" type="string" default="ghcr.io/usestrix/strix-sandbox:0.1.12">
  Docker image to use for the Strix sandbox environment.

  You can override this to use a custom sandbox image or a specific version.

  ```bash theme={null}
  export STRIX_IMAGE="ghcr.io/usestrix/strix-sandbox:0.1.12"
  ```
</ParamField>

<ParamField path="STRIX_RUNTIME_BACKEND" type="string" default="docker">
  Runtime backend for Strix sandboxes.

  Currently, only `docker` is supported.

  ```bash theme={null}
  export STRIX_RUNTIME_BACKEND="docker"
  ```
</ParamField>

<ParamField path="STRIX_SANDBOX_EXECUTION_TIMEOUT" type="string" default="120">
  Timeout in seconds for sandbox tool execution.

  This controls how long Strix will wait for individual tool executions (terminal commands, Python scripts, etc.) to complete.

  ```bash theme={null}
  export STRIX_SANDBOX_EXECUTION_TIMEOUT="120"
  ```
</ParamField>

<ParamField path="STRIX_SANDBOX_CONNECT_TIMEOUT" type="string" default="10">
  Timeout in seconds for connecting to the sandbox tool server.

  ```bash theme={null}
  export STRIX_SANDBOX_CONNECT_TIMEOUT="10"
  ```
</ParamField>

<ParamField path="DOCKER_HOST" type="string">
  Docker daemon connection URL.

  By default, Strix uses the local Docker daemon. Set this to connect to a remote Docker host.

  ```bash theme={null}
  export DOCKER_HOST="tcp://remote-host:2375"
  ```
</ParamField>

## Telemetry

<ParamField path="STRIX_TELEMETRY" type="string" default="1">
  Enable or disable anonymous telemetry.

  Set to `0` to opt out of telemetry collection. Telemetry helps improve Strix by collecting anonymous usage statistics.

  ```bash theme={null}
  export STRIX_TELEMETRY="0"
  ```
</ParamField>

## Example Configurations

### OpenAI GPT-5

```bash theme={null}
export STRIX_LLM="openai/gpt-5"
export LLM_API_KEY="sk-..."
```

### Anthropic Claude

```bash theme={null}
export STRIX_LLM="anthropic/claude-sonnet-4-6"
export LLM_API_KEY="sk-ant-..."
```

### Strix Router (Multiple Providers)

```bash theme={null}
export STRIX_LLM="strix/gpt-5"
export LLM_API_KEY="your-strix-router-key"
```

### Ollama (Local)

```bash theme={null}
export STRIX_LLM="ollama/llama3.1:70b"
export LLM_API_BASE="http://localhost:11434"
```

### With Web Search

```bash theme={null}
export STRIX_LLM="openai/gpt-5"
export LLM_API_KEY="sk-..."
export PERPLEXITY_API_KEY="pplx-..."
```

## Configuration Persistence

Strix automatically saves your environment variables to `~/.strix/cli-config.json`. This means you only need to set these variables once, and Strix will remember them for future runs.

To override saved configuration temporarily, you can:

1. Set environment variables in your current shell session
2. Use a custom config file with the `--config` flag
3. Clear a variable by setting it to an empty string: `export STRIX_LLM=""`

<Note>
  LLM configuration variables (`STRIX_LLM`, `LLM_API_KEY`, etc.) are re-validated on each run. If you change these in your environment, Strix will use the new values and update the saved configuration.
</Note>