What does peek do that `ls` or `cloc` does not?

It answers the question you actually have about an unfamiliar repo: what is this, which language dominates, where does the code live, and which files should I read first — in one screen instead of five commands.

Does it send my code anywhere?

No. Everything runs locally and works offline. The optional `--llm` summary is the only feature that talks to an API, and it stays off unless you set a key yourself.

Yes — the repo keeps a set of issues labelled good first issue. Install with pip install -e "./peek[dev]" and the test suite runs locally.

Hariom Lohar — peek · On PyPI · shipped

A zero-config codebase cartographer. Point it at any repository and it draws the map: languages, size, structure and the files that actually matter — in about five seconds, entirely offline.

pip install peek-code
peek .

The default is an interactive TUI built on Rich and Textual — j/k to move, / to filter, o to open a file in $EDITOR, enter for details. peek --no-tui prints a static Rich report for screenshots, pipes and CI, and peek --html -o map.html writes a shareable page.

Two commands go further than a summary. peek --pack selects the highest-signal files within a token budget and prints them as one LLM-ready bundle — peek --pack --ask "auth" narrows it to a topic. peek trace walks a function tree across files: what calls it, what it calls, and how deep, with --html for a browsable version.

peek find "token validation" . --limit 10
peek trace scan --depth 3 --html

Python 3.11+, ten themes, 146 tests, MIT. Issues tagged good first issue are open if you want to build on it.

Highlights

Read any codebase in about five seconds, fully offline
Interactive TUI, static Rich output, or a shareable HTML map
`--pack` builds an LLM-ready bundle inside a token budget
`peek find` ranks matches across filenames and content
`peek trace` walks the function tree across files, with an HTML view
Ten themes, 146 passing tests, zero configuration

CLI TUI Rich + Textual Python PyPI developer tools

Python 89.2% · HTML 10.8%

Tests

146

Themes

Python

3.11+