# IndieML - Substance API

A lightweight, transformer-based API that evaluates text for substance, depth, and clarity in under 25ms while achieving 87% of the accuracy of a full-scale LLM. 
This package provides REST API, Python SDK, and MCP configuration details to natively connect the hosted Substance API to a wide variety of applications.
For more information, including API key requests, please visit https://indieml.app.

## REST API 

For integrations outside the Python and MCP ecosystems (such as standard OpenAI function calling, custom AI scripts, or raw HTTP requests), the hosted endpoint accepts standard JSON payload requests.

**cURL**
```bash
curl -X POST https://indieml.app/v1/score/substance \
  -H "X-API-Key: YOUR_API_KEY_HERE" \
  -H "Content-Type: application/json" \
  -d '{"input_text": "This is a test run."}'
```

**Python (Requests)**
```python
import requests

response = requests.post(
    "https://indieml.app/v1/score/substance",
    headers={"X-API-Key": "YOUR_API_KEY_HERE"},
    json={"input_text": "This is a test run."}
)
print(response.json())
```

**Node.js (Fetch)**
```javascript
const response = await fetch("https://indieml.app/v1/score/substance", {
  method: "POST",
  headers: {
    "X-API-Key": "YOUR_API_KEY_HERE",
    "Content-Type": "application/json"
  },
  body: JSON.stringify({ input_text: "This is a test run." })
});
console.log(await response.json());
```

## Python SDK

**Installation** (requires Python 3.11 or higher)
```bash
pip install indieml
```

**Usage**
```python
from indieml import Substance

# Initialize the API
api = Substance(api_key="YOUR_API_KEY_HERE")

# Evaluate text for substance, depth, and clarity
result = api.score("This is a test run.")
print(result)
```

## MCP Server Integration

You can integrate the Substance API into AI coding assistants by routing them to our cloud-native ASGI endpoints.

### Cursor, ChatGPT, and Claude Code (Streamable HTTP)
For agents that support native Streamable HTTP / SSE configurations, provide the remote URL and authorization header directly:
```json
{
  "mcpServers": {
    "indieml": {
      "type": "http",
      "url": "https://indieml.app/mcp/",
      "headers": {
        "X-API-Key": "YOUR_API_KEY_HERE"
      }
    }
  }
}
```

### Claude Desktop (stdio via npx Bridge)
For local Linux, Mac, and Windows clients that require stdio transport, use the npx mcp-remote bridge to route the connection to the cloud endpoint. Node.js v20+ is required. Ensure your OS environment variables include INDIEML_API_KEY.
```json
{
  "mcpServers": {
    "indieml": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://indieml.app/mcp/",
        "--header",
        "X-Api-Key:${INDIEML_API_KEY}"
      ],
      "env": {
        "INDIEML_API_KEY": "YOUR_API_KEY_HERE"
      }
    }
  }
}
```

## Applications & Use Cases

The Substance API is a fast and inexpensive pre-screening tool designed to integrate with your existing classification pipeline.

While it evaluates substance across any text, it is uniquely suited for high-volume, automated data pipelines:

### 1. RAG & AI Knowledge-Base Preprocessing
Pre-filter text before expensive LLM vectorization and processing.
* **Knowledge Portals:** Filter out bloated, low-signal documentation prior to indexing.
* **Document Archives:** Rank corporate or research archives based on actual information density.
* **Feed Aggregation:** Rank news, newsletters, or educational content by structural depth rather than just recency or click-through rates.

### 2. Automated Content Triage & Large-Scale Screening
Route inbound text based on substance, clarity, and depth rather than relying solely on keywords or sentiment analysis.
* **Support Tickets:** Identify highly detailed bug reports and instantly route them past basic classification bots to human operators.
* **Customer Feedback:** Isolate high-quality product critiques from generic "it's great" or "it's broken" noise.
* **User-Generated Content:** Surface high-effort, high-value forum posts and survey responses while automatically burying low-effort spam.

<!-- mcp-name: io.github.ezchx/indieml -->