Extracting data from equipment technical manuals
Equipment manufacturers publish a technical manual for each machine line, each with its own section numbering and table styles. Dealer service departments and fleet maintenance teams keep one for every model they support, and a single manual runs to hundreds of pages. They arrive as portal downloads and scans of printed binders.
Maintenance planning needs the specifications a technician looks up: fastener torque by size and grade, fluid specifications and fitting torques. A CMMS loads those values into the task list for each asset, so every work order carries the specification for the job.
What makes equipment technical manuals hard
Values defined by stacked headers
A torque figure means something only under its full header stack: grade or property class, lubricated or dry, then N·m or lb-ft.
Diagrams carry the key
Head and nut markings drawn above the table tell a technician which grade column applies to the fastener in hand.
Footnotes qualify the headers
Footnote markers on the Lubricated and Dry headers point to definitions printed below the table, which set the condition each column assumes.
Parse, extract, verify
Parse
Parse turns the document into reading-order Markdown and a hierarchy of blocks (text, tables, figures), and grounds each block to its exact location on the page.
Extract
You define the fields in a JSON schema, and Extract returns their values in a consistent, structured format, each linked back to where it appears in the Markdown.
Verify
Each value links back to where it appears, down to the character range in the Markdown, and the page and bounding box in the source document, so every result can be verified.
One equipment technical manual, end to end
About this document: Five portrait pages from a manufacturer's technical manual for a line of compact utility tractors, taken from its repair information group. Fastener torque tables, metric then inch, are followed by O-ring fitting service pages and a diesel fuel specification. Page 2, the inch fastener table, is featured. The pages hold reference data only and are shown unredacted.
Processing objective: The objective is to read one complete torque lookup: a bolt size and the four values one SAE grade takes, lubricated and dry, in newton-meters and pound-feet. Each figure is correct only when paired with its size row and every header tier above it.
Demonstration fields: Bolt size (in), SAE grade 5 lubricated torque (N·m), SAE grade 5 lubricated torque (lb-ft), SAE grade 5 dry torque (N·m), SAE grade 5 dry torque (lb-ft)
The document
4 of 5 pages
Page 1
Page 2shown below
Page 3
Page 4
Where each value came from
inch_fastener_torque.rows[4].size
{
"value": "1/2",
"page": 2,
"range": {
"end": 8393,
"start": 8390
},
"box": {
"xmin": 0.07297,
"ymin": 0.47545,
"xmax": 0.12217,
"ymax": 0.49828
}
}inch_fastener_torque.rows[4].grade_5.lubricated_nm
{
"value": 85,
"page": 2,
"range": {
"end": 8492,
"start": 8490
},
"box": {
"xmin": 0.50999,
"ymin": 0.47545,
"xmax": 0.55821,
"ymax": 0.49826
}
}inch_fastener_torque.rows[4].grade_5.lubricated_lb_ft
{
"value": 63,
"page": 2,
"range": {
"end": 8503,
"start": 8501
},
"box": {
"xmin": 0.55821,
"ymin": 0.47545,
"xmax": 0.60645,
"ymax": 0.49826
}
}inch_fastener_torque.rows[4].grade_5.dry_nm
{
"value": 110,
"page": 2,
"range": {
"end": 8515,
"start": 8512
},
"box": {
"xmin": 0.60645,
"ymin": 0.47545,
"xmax": 0.65526,
"ymax": 0.49826
}
}inch_fastener_torque.rows[4].grade_5.dry_lb_ft
{
"value": 80,
"page": 2,
"range": {
"end": 8526,
"start": 8524
},
"box": {
"xmin": 0.65525,
"ymin": 0.47545,
"xmax": 0.70349,
"ymax": 0.49826
}
}How to try it yourself
Everything on this page is reproducible. The sample, its schema, and the Parse and Extract API outputs are published. You can repeat the same two calls from whichever surface you already build on.
- Claude Skilldescribe the job; the agent writes it
- CLIno code at all
- MCPexpose ADE docs as an agent tool
- REST APIany language
- Pythonlandingai-ade
- TypeScriptlandingai-ade
Documents into data
Accurate
The best accuracy on the difficult documents other tools get wrong, such as messy scans, complicated tables, checkboxes and handwriting.
Auditable
Every answer cites its sources. Values come back with an audit trail (page, bounding box, character range) so a human can review them.
Agentic
Applies multi-step reasoning, and plans before it acts. Handles documents it has never seen before and adapts to variations.
API-first
Composable APIs that drop into any tech stack and can be called from anywhere. Easy to build on and embed in your pipelines and products.
Affordable
Cost-competitive rates with complete price transparency. Spend controls before you run, and every job reports its exact cost.
Enterprise-grade
Built for production and scale from day one. Fully observable and monitored, running securely in our cloud or yours.
Built for high-volume, high-throughput production
- Zero data retention
- Priority rate limits
- Enhanced support
- VPC and on-prem deployment
Available on Team and Enterprise plans. VPC and on-prem deployment require an Enterprise plan.
