ade-cli
Command-line tool to parse complex documents into Markdown and extract schema-shaped fields with bounding box evidence.
Parse policy forms, claims, and invoices into JSON with bounding-box visual evidence directly from the terminal.
ade-cli is a standalone command-line interface for LandingAI's Agentic Document Extraction service. It converts complex insurance PDFs, tables, and policy forms into structured JSON output accompanied by visual bounding boxes and grounded Markdown. It is designed for developers building automated document processing pipelines and AI agent workflows.
What it is
The tool splits extraction into two main commands: `ade parse` converts files into grounded Markdown and structural elements, while `ade extract` applies custom JSON schemas to pull specific fields. Built-in inspection features like `ade view` open local interactive HTML reports showing exactly where each extracted field originates on the original page.
Who it's for
It suits insurance technical teams and AI engineers who need to process policy applications, claims, and complex invoices via scripts or CLI agents. It is less suitable for non-technical teams looking for a standalone web GUI, as it requires comfort with terminal commands and custom JSON schema design.
Key features
Schema-driven extraction
Pulls JSON data according to your custom schema files with page-and-box evidence for every extracted value.
Local result store
Saves job outputs locally in `~/.ade` so re-running identical extractions consumes no additional API credits.
Agent-first design
Provides a single `ade help --json` call exposing all commands, flags, and exit states for automated AI agents.
Visual evidence viewing
Generates self-contained HTML reports and crops specific document regions to PNG for human verification.
Up and running in 3 steps
Our honest verdict
What's great
- Runs via self-contained binaries on macOS, Linux, and Windows without needing a Python runtime.
- Emits clean, predictable JSON objects on stdout for direct scripting in CI and agent workflows.
- Avoids duplicate API charges by storing parsed document elements locally in `~/.ade`.
- Includes built-in interactive HTML viewer (`ade view`) to visually verify extracted data locations.
Where it falls short
- Requires an external LandingAI account and API key, which incurs usage-based credit costs for document parsing.
- Does not offer a complete self-hosted extraction engine; the CLI acts as a client for LandingAI cloud processing.
- CLI-only interface requires terminal expertise and manual JSON schema creation.
Pricing & license
The CLI binary is completely free and open-source under the Apache-2.0 license. However, parsing documents requires a LandingAI API key which consumes service credits based on document processing usage.
Open remote AI-training roles
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Do I need to install Python to use ade-cli?
No, each release includes self-contained compiled binaries for macOS, Linux, and Windows.
How does ade-cli prevent unnecessary credit usage?
Parsed document elements are saved locally in your `~/.ade` folder, making identical re-runs free of API credit charges.
Can AI agents execute ade-cli directly?
Yes, running `ade help --json` outputs the complete command surface and schemas in JSON, adhering to the included `SKILL.md` contract.
