Finn Clancy
← Selected work
Project

finnslib

Summarise repos for LLM context

Python package to summarise a function, file, or whole repo tree for LLM prompts — cheap gpt-4o-mini passes, CLI included.

View on PyPI ↗
01 · Type
Project
02 · Year
2025
03 · Status
Published

What it is

finnslib helps you paste less into LLM prompts. Summarise a function, a file, or a whole repo tree instead of dumping raw source.

Problem

Pasting {filename}\n{filecontent} for ten files burns tokens and time.

Python API (examples)

  • summarise_file() — what the file does.
  • summarise_function() — what a single function does.
  • write_repo_tree() — .txt tree on disk.
  • summarise_repo_tree() — tree + per-file blurbs.
  • Detail knob: "short" | "medium" | "long".

Optional OpenAI key; uses cheap gpt-4o-mini so a full tree pass is usually well under a cent.

CLI

  • finn_tree . — emit structure.
  • finn_file_sum main.py — file summary.
  • finn_func_sum file.py func_name — function summary.
  • finn_tree_sum . — tree + per-file summaries.

Who it is for

  • Onboarding to large OSS repos.
  • Long files you do not want to re-read end-to-end.
  • Teams standardising LLM context packs.