Currently api-squash is CLI-only. A public Python API would allow programmatic use:
```python
from api_squash import summarize_project, summarize_file
result = summarize_project("src/", no_docstrings=True)
```
The internals already support this — `extractor.py`, `renderer.py`, and `scanner.py` do the work. This is a matter of defining the public API surface, stabilizing it, and documenting it.
This also unblocks:
- MCP server integration (call api-squash from an MCP tool handler)
- GitHub Action (invoke from Python in a workflow)
- Embedding in other tools
Currently api-squash is CLI-only. A public Python API would allow programmatic use:
```python
from api_squash import summarize_project, summarize_file
result = summarize_project("src/", no_docstrings=True)
```
The internals already support this — `extractor.py`, `renderer.py`, and `scanner.py` do the work. This is a matter of defining the public API surface, stabilizing it, and documenting it.
This also unblocks: