Capstone — Python CLI That Summarizes With Claude

From Zero to Useful Tool

A good capstone is a script you'll actually use. This one reads any text file and streams a Claude summary to your terminal, with sensible errors and retries.

pip install anthropic export ANTHROPIC_API_KEY=sk-ant-One dependency, one env var. That is the setup.

# summarize.py import argparse import sys import time from pathlib import Path import anthropic def read_text(path: Path) -> str: if not path.exists(): sys.exit(f"error: {path} does not exist") if path.stat().st_size > 500_000: sys.exit(f"error: {path} is larger than 500 KB") return path.read_text(encoding="utf-8") def summarize(text: str, bullets: int, tries: int = 3) -> None: client = anthropic.Anthropic() prompt = ( f"Summarize the following text in exactly {bullets} bullet points. " f"Each bullet under 20 words.\n\n{text}" ) for attempt in range(1, tries + 1): try: with client.messages.stream( model="claude-opus-4-7", max_tokens=1024, messages=[{"role": "user", "content": prompt}], ) as stream: for chunk in stream.text_stream: print(chunk, end="", flush=True) print() return except anthropic.APIStatusError as e: if e.status_code in (429, 529) and attempt < tries: wait = 2 ** attempt print(f"\n[retry in {wait}s: {e.status_code}]", file=sys.stderr) time.sleep(wait) continue raise def main() -> None: p = argparse.ArgumentParser(description="Summarize a text file with Claude.") p.add_argument("file", type=Path) p.add_argument("--bullets", type=int, default=5) args = p.parse_args() text = read_text(args.file) summarize(text, bullets=args.bullets) if __name__ == "__main__": main()argparse for args, streaming for UX, exponential backoff on 429/529, size guard on input. Production-ish in 40 lines.

# usage python summarize.py README.md --bullets 3 python summarize.py meeting-notes.txt --bullets 7One script, many files. The kind of tool you keep on your PATH.

The big idea: programming with AI is still programming. The skills you built across this track — types, tests, error handling, prompting — all show up in one small script that actually does something useful.