Metadata-Version: 2.4
Name: gutenberg2-sdk
Version: 0.1.0
Summary: Python SDK + MCP server for Gutenberg: datasets of AI runs, SQL, LLM/SAE evaluation workflows.
Requires-Python: >=3.12
Requires-Dist: httpx>=0.27
Requires-Dist: mcp>=1.0
Requires-Dist: pydantic>=2.7
Requires-Dist: python-dotenv>=1.0
Description-Content-Type: text/markdown

# gutenberg2-sdk

Python SDK and read-only MCP server for [Gutenberg](https://console.gutenberg.ai) —
datasets of AI runs, read-only SQL, LLM/SAE evaluation workflows, and reports.

```bash
pip install gutenberg2-sdk
export GUTENBERG_API_KEY=gtn2_...   # create one at https://console.gutenberg.ai/settings
```

```python
from gutenberg2 import Gutenberg

client = Gutenberg()
dataset_id = client.create_dataset(name="my runs")
client.add_runs(dataset_id, [...])
rows = client.execute_sql(dataset_id, "SELECT id FROM runs")
```

Optional configuration: `GUTENBERG_BASE_URL` (defaults to
`https://api.gutenberg.ai`), `GUTENBERG_DATASET_ID`, `GUTENBERG_MAX_INPUT_TOKENS`.
Run the MCP server with `gutenberg2 mcp`.
