Ladex for VS Code
Flags AI components inline as you edit Python — inference APIs, models, agent frameworks,
vector stores — and nudges what each obligates under the EU AI Act. Nothing shows on non-AI
code (ruthless silence).
All detection runs in the Ladex Python engine, the same engine the CLI and PR check use —
the editor can never disagree with the gate.
Install
The platform builds bundle the Ladex engine — one install, no separate pip install.
- From the Marketplace: search “Ladex” in the Extensions panel (or open the
listing) — VS Code
automatically installs the build for your OS/CPU.
- From a
.vsix: download the one matching your platform from the
Releases page —
ladex-<version>-{win32-x64,darwin-arm64,linux-x64}.vsix — then Extensions panel → ⋯ →
Install from VSIX… (or code --install-extension <file>.vsix).
Open any Python file that uses an AI library — diagnostics appear as you type. No engine setup.
On a platform without a bundled build (e.g. Intel Mac, linux-arm64), install
ladex-<version>-universal.vsix and provide the engine yourself: pip install ladex.
Configuration
| Setting |
Default |
Meaning |
ladex.serverCommand |
ladex |
Overrides the engine command. Leave as ladex to use the bundled binary (falling back to a ladex on PATH); set a full path to force a specific engine. |
ladex.serverArgs |
["serve"] |
Arguments passed to the engine command. |
The extension resolves the engine in this order: an explicit non-default serverCommand →
the bundled binary (bin/ladex) → a ladex on your PATH. If it can't start any, it shows
a message telling you to pip install ladex or set ladex.serverCommand.
Development (contributing to the extension)
You only need this if you're hacking on the extension itself:
cd extensions/vscode
npm install
npm run compile # or: npm run watch
Then press F5 — the bundled launch config compiles, puts the repo's .venv on PATH, and
opens demo/ in an Extension Development Host so you can see live diagnostics.
Build a distributable package with npm run package → ladex-<version>.vsix.
