LimAI Docs

$limai vision <subcommand>

Run vision queries against PDFs and images via Gemini.

Options

--jsonOutput raw JSON

--verboseEnable debug logging

Subcommands

Ask a vision question about one or more PDFs or images. Returns plain text by default, or structured output when a JSON Schema file is provided.

$limai vision ask

Ask a vision question about one or more PDFs or images.

Options

-f, --files <path>File path (repeatable) — PDFs or images (required)

-p, --prompt <text>The question to ask (required)

--project <id>Project ID (defaults to LIMAI_PROJECT_ID env var)

--schema-file <path>JSON Schema file — when provided, returns structured output

--fastUse the faster Gemini 3.1 Flash Lite model instead of Gemini 3.1 Pro

limai vision ask --project proj_abc123 -f ./invoice.pdf -p "What is the total amount due?"

limai vision ask --project proj_abc123 -f ./page1.png -f ./page2.png -p "Do these pages belong to the same document?"

limai vision ask --project proj_abc123 -f ./invoice.pdf -p "Extract the invoice header fields" --schema-file ./invoice-schema.json

limai vision ask --project proj_abc123 -f ./receipt.jpg -p "Is this a receipt or an invoice?" --fast