Skill Detail

Extract schema-backed document data with ADE CLI

Use LandingAI's ADE CLI to parse visually complex documents, cache parse artifacts, and extract schema-shaped fields with page and bounding-box evidence from an agent or CI workflow.

Data Extraction & TransformationMulti-Framework
Data Extraction & Transformation Multi-Framework Security Reviewed
⭐ 2.4k GitHub stars
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill extract-schema-backed-document-data-with-ade-cli Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
ADE CLI, LandingAI ADE API key, source documents or document URLs, optional jq or CI shell tooling
Install & setup
Download the platform-specific ADE CLI release from https://github.com/landing-ai/ade-cli/releases or review and run the upstream installer from the repository. Authenticate with ade auth login or set ADE_API_KEY for automation.
Author
LandingAI
Publisher
Company Maintained Open Source
Last updated
Aug 1, 2026
Quick brief

# Extract schema-backed document data with ADE CLI

How it works

What this skill actually does

Use this skill when an agent or operator needs repeatable document parsing and structured extraction from PDFs or other visually complex documents, with local artifact reuse and evidence-backed output. ADE CLI wraps LandingAI’s Agentic Document Extraction API from the terminal so parse jobs, markdown, elements, extraction JSON, provenance, and history can be composed in shell, CI, cron, or an agent harness.

## What the agent/operator does

1. Install the ADE CLI and configure an ADE API key with ade auth login or ADE_API_KEY.
2. Parse a document with ade parse -d document.pdf –json, optionally including markdown or elements in the JSON output.
3. Reuse the resulting job item id for downstream work instead of reparsing unchanged input.
4. Run ade extract JOB_ID –schema schema.json –json to produce schema-shaped extraction output.
5. Inspect per-field evidence, page references, and bounding boxes when validating results.
6. Use ade history and the local store to resume pending jobs, audit prior parses, or clear stale artifacts.

## When to invoke

Invoke this instead of uploading documents through a web product when extraction needs to be scripted, repeated across a document batch, embedded in CI, cron, or agent harnesses, or verified with machine-readable evidence. It is strongest for invoices, forms, reports, tables, figures, charts, and other visually complex documents where markdown plus structured fields need to stay tied to source locations.

## Scope boundary

This is not a general OCR category, document-management product, or SDK listing. Keep the workflow scoped to terminal-driven ADE parse and extract jobs through ade, with explicit schema input, local artifact reuse, JSON output, and evidence review.

## Requirements

– ADE CLI installed on macOS, Linux, or Windows.
– A LandingAI ADE API key.
– Source documents or document URLs.
– A JSON schema when running structured extraction.

## Installation

Download the platform-specific ADE CLI release from the upstream GitHub releases page or review and run the upstream installer from the repository. After installation, run ade auth login or set ADE_API_KEY for non-interactive automation.

## Verification notes

Source metadata was verified from the exact GitHub repository and README. As of this intake run, landing-ai/ade-cli had 2,395 GitHub stars, Apache-2.0 license metadata, push activity on 2026-07-31 UTC, official LandingAI docs, and README evidence for ade parse, ade extract, stable –json output, local parse and extract artifacts, environment selection, history, and evidence-backed extraction.