LEX AI Command Reference
The five lex ai-* commands and their purpose, the hyphenated setup command, why ai-verify runs silently, and what the issue-report bundle contains.
LEX AI Command Reference
This page summarizes the user-facing LEX AI commands and their purpose. Run them from the project root. All of them are hyphenated.
Core commands
lex ai-dashboard: open the configuration dashboard for mode selection, coding-environment selection, project state, and credentials. It binds to127.0.0.1on a random port and opens your browser.lex ai-faq: render the FAQ page locally from the installed package and open it at the Prompt Builder. It takes no options and fetches nothing from the internet, so it works offline.lex ai-issue-report: build a support bundle under.lex-ai-reports/, and optionally submit it through the remote MCP when one is configured.lex ai-update: upgrade the package and run the re-runnable migration steps for an existing setup, refreshing the copied assets and the MCP registration.lex ai-verify: make the on-disk assets match what the installed packages ship, restoring drifted files and pruning stale mode assets that would otherwise leave a mixed tool surface.
Setup command relationship
lex setup-with-ai wires the local MCP server into the current project and writes the IDE configuration. Run it once per project, before the ai-* commands. The spelling is hyphenated: lex setup_with_ai is a Python module path, not a CLI command, and fails with No such command.
Why ai-verify is quiet
lex ai-verify is also the MCP pre-flight, run as a subprocess at the start of every tool call. That is why --silent must print nothing on success — the server speaks JSON-RPC over stdout, and any stray output would corrupt the stream. Write failures are still reported even under --quiet and --silent.
About the issue-report bundle
Credential values are masked on the way in: every collected file is redacted before it is written, and the bundle's manifest.json records which files were touched and how many values were masked. The bundle still carries local IDE and MCP configuration, so review it before sharing. The ZIP is always written first; ticket submission is best-effort and never blocks it.
Prompt starter
A reliable starter pattern:
I want to [build / edit / review / document / test / deploy] a LEX app for [rough idea].
I selected [mode] in the LEX AI dashboard.
Please ask me for the missing details that matter.
If the rough idea will not fit in that blank yet, select Project Brief instead and let the interview produce the contract. See LEX AI commands and Starter prompts.
Was this helpful?
Your feedback shapes what we write next.
