The dashboard

las dashboard ~/code                  # terminal table
las dashboard ~/code --html           # + ~/.localalibi/dashboard.html
las dashboard ~/code --lang python    # only projects with Python files
las dashboard ~/code --depth 3        # look deeper
las dashboard ~/code --no-cache       # re-analyse everything

Columns

Column	Content
Project	path relative to the root; ≈ for a probable project
Type	project type: Python, Node.js/TypeScript, Monorepo (Python + Node.js/TypeScript), ≈Python scripts…
Stack	frameworks (≈ when only declared), or the main language
DB	databases (≈ when only declared)
Services	external services: OpenAI, Anthropic, Stripe, AWS, Supabase…
Changed	last change: the newest of the last commit and the newest file (git when it came from git)
Since last	what changed since the previous analysis: +2 -1 ~3 ⚠1, — (nothing, served from cache) or new
Alerts	⚠ contradictions and ? unknowns

Rows are sorted by last change, newest first.

How projects are found

The dashboard uses exactly the same rules as las check: it walks down, takes every folder with .git or a manifest as a project and does not look inside it (a monorepo counts once), and when a folder contains no project at all, it asks for its verdict — loose script folders become ≈ probable projects, and deeper folders of projects are expanded.

Hidden folders, node_modules, .venv and any virtualenv (a folder with pyvenv.cfg, whatever its name) are skipped.

If you point it at a folder that is itself a project, it shows that project's profile and suggests las scan.

Cache

Each project's profile is saved in its .localalibi/project.json together with a fingerprint: the path, modification time and size of every file, the git HEAD, and the tool version. If the fingerprint matches, the profile is reused; otherwise the project is analysed again. On a folder with 50+ real projects, a cached run takes about a second.

The HTML report

--html writes ~/.localalibi/dashboard.html: one self-contained file (styles, script and icons inline, no network), with search, a language filter and sorting. Each project expands to its full profile with every piece of evidence. Paths under your home folder are shown as ~/…, so the file is safe to share.

Open the example report — Express, Flask, httpx, Vite, the FastAPI full-stack template and others.