Quickstart
One image, or many in a single warm process — a single Reader is reused across every image passed
to sceptre run:
sceptre run receipt.png --lang english --format jsonsceptre run page1.png page2.png page3.png # batch: models load oncesceptre run sign.jpg --lang english --lang korean # multi-languagesceptre run page.png --lang english --timings # per-stage load/detect/recognize breakdownsceptre run huge-scan.png --canvas-size 1600 # trade some accuracy for lower peak memory--format accepts text (the default, tab-separated) or json. --no-detail emits only the
recognized text, omitting confidence and box detail. See the full CLI reference for
every flag and subcommand.
Decodes PNG, JPEG, BMP, GIF, TIFF, WebP, and NetPBM — all pure-Rust decoders. See Image formats for what’s out of scope.
Library
Section titled “Library”use sceptre::{Reader, ReadOptions};
let reader = Reader::builder().build()?;for line in reader.readtext("receipt.png".as_ref(), &ReadOptions::default())?.lines { println!("{} ({:.2})", line.text, line.confidence);}# Ok::<(), sceptre::OcrError>(())This requires a backend feature and download enabled in Cargo.toml — see
Installation. Full walkthrough in the Library guide.
MCP server
Section titled “MCP server”sceptre mcp --lang englishStarts the stdio MCP server (requires the CLI’s mcp feature) with a readtext tool that runs OCR
over an image file and returns recognized lines with their bounding boxes and confidences. See the
MCP server guide for the tool schema and an example invocation.