Local OCR

What it does

When your agent uses it

• Needing to read small/dense text in screenshots
• Documents that a multimodal model may misread
• Exact text plus bounding box coordinates are needed for field extraction
• Layout analysis

Example prompts

• “/local-ocr”

Requirements

• Python 3
• Node.js

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Never fabricate OCR text. Only use the text field from results. If confidence < 0.5, state it.
2. 2Cite coordinates when relevant. Box coordinates are in pageSpace.
3. 3Use --schema-version 1 for reproducible output. Do not parse help text.
4. 4Prefer detect first on large images, then recognize --region on areas of interest.
5. 5Check exit codes before parsing stdout. Non-zero exit means stdout may be empty; read stderr.

What it can do on your machine

Read from SKILL.md and the folder at commit d2d558c. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • python

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Local OCR loads about 1.4k tokens when it runs. Until then it costs about 119 tokens; SKILL.md has 462 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~119

When it runs · the whole SKILL.md, loaded when a task matches

~1.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from arcships/light-ocr at commit d2d558c, republished under its Apache-2.0 licence (© arcships). 462 words, ~1,415 tokens.

name

local-ocr

description

Extract text from local images (PNG, JPEG) with precise coordinates, confidence scores, and stable error handling using the light-ocr CLI. Use when needing to read small/dense text in screenshots, receipts, labels, forms, or documents that a multimodal model may misread; when exact text plus bounding box coordinates are needed for field extraction, redaction, counting, or layout analysis; or when deterministic offline OCR is required without network or API calls.

# Step 1: recognize the full image
light-ocr screenshot.png --format json

# Step 2: re-run on the specific region (coordinates from step 1 boxes)
light-ocr recognize screenshot.png --region 100,80,640,320 --format json

# Step 1: detect where text regions are (fast, no recognition)
light-ocr detect receipt.png

# Step 2: recognize only the region containing the target field
light-ocr recognize receipt.png --region 50,200,300,80 --format json

light-ocr detect image.png | python -c "import json,sys; print(len(json.load(sys.stdin)['pages'][0]['detections']))"

light-ocr image.png --format text

for f in *.png; do
  light-ocr recognize "$f" --format jsonl
done

# Full PDF with JSON output
light-ocr report.pdf --format json

# Page range with streaming JSONL
light-ocr report.pdf --pages 1-10 --format jsonl

# Multiple images as one document
light-ocr document scan1.png scan2.png --format text

# Check if PDF support is available
light-ocr document info

light-ocr doctor --json

Need text from an image or document?
├── Know which region? → recognize --region x,y,w,h --format json
├── Need full text only? → recognize --format text
├── Need text + coordinates? → recognize --format json
├── Only need where text is? → detect
├── Large image, unsure where text is? → detect first, then recognize --region
├── PDF? → light-ocr <file.pdf> --format json
├── Multiple images? → light-ocr document <files> --format json
├── Need system/hardware info? → doctor --json
└── Need engine info or version? → info --model-info / info --version

{
  "schemaVersion": 1,
  "source": { "kind": "image", "mediaType": "...", "identity": {}, "appliedTransforms": {} },
  "pages": [{
    "index": 0,
    "width": 640, "height": 480,
    "coordinateSpace": "pageSpace",
    "structure": "ocr-order",
    "lines": [{ "id": "L0", "text": "HELLO", "confidence": 0.99, "box": [4 points] }]
  }]
}

© arcships, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in .agents/skills/local-ocr of arcships/light-ocr.

• SKILL.md
• agents/openai.yaml

Open the folder on GitHubat commit d2d558c