# Disco
> Not another AI data analyst. A discovery pipeline that finds novel, statistically validated patterns in tabular data — feature interactions, subgroup effects, and conditional relationships you wouldn't think to look for.

## What It Does
- Finds complex patterns — interactions, thresholds, subgroup effects — without requiring prior hypotheses
- Validates every pattern on hold-out data with FDR-corrected p-values
- Checks each finding against academic literature for novelty, with citations
- Returns structured, machine-readable results agents can reason over directly

## Links
- SDK: https://pypi.org/project/discovery-engine-api/
- Full docs (LLM-friendly): https://disco.leap-labs.com/llms-full.txt
- MCP server: https://disco.leap-labs.com/.well-known/mcp.json
- API spec: https://disco.leap-labs.com/.well-known/openapi.json
- Visualization spec: https://disco.leap-labs.com/visualization-spec
- API keys: https://disco.leap-labs.com/docs
- Interactive reports: https://disco.leap-labs.com/discover

## Getting an API Key
**Programmatic (for agents):** Two-step signup — `POST /api/signup` sends a verification code to the email, then `POST /api/signup/verify` submits the code and returns a `disco_` API key. No auth required.
**Manual (for humans):** Sign up at https://disco.leap-labs.com/sign-up, create key at https://disco.leap-labs.com/docs.
Free tier active immediately (10 credits/month, unlimited public runs). No credit card required.

## Quick Start
```bash
pip install discovery-engine-api
```

```python
from discovery import Engine

engine = Engine(api_key="disco_...")
result = await engine.discover(file="data.csv", target_column="outcome")

for pattern in result.patterns:
    if pattern.p_value < 0.05 and pattern.novelty_type == "novel":
        print(f"{pattern.description} (p={pattern.p_value:.4f})")
```

## Cost
- Public runs: Free (results published, depth=1)
- Private runs: 1 credit/MB/depth × 5 with LLM explanations ($0.10/credit)
- Free tier: 10 credits/month