Process an entire document collection
Connect a documentation site, object store, or folder of files. mGenie identifies document classes, links revisions, and extracts facts, procedures, warnings, and relationships.
01 — Import
One URL or a folder of PDFs. mGenie found 2,481 documents and 314 products at ACME in 12 minutes.
02 — Classify and extract
Specifications, compatibility rules, procedures, and warnings are extracted with their conditions and effective dates.
03 — Query
client.compatibility.check({
product: "MX900",
component: "B14"
})
Your portal, service app, and support tooling all read from the same current model.
CUSTOM DOCUMENT ENGINE
Improve extraction with reviewed corrections
When reviewers correct a document class, field, relationship, or table, mGenie applies that feedback to similar files in the same corpus. Teams work with familiar labels and schemas instead of configuring model training.
Track changes across revisions
When a bulletin or release note arrives, mGenie records each changed value with its date and source. Review the changes in the app or receive them through a webhook.
Today, 14:32
Service Bulletin #143 added
July 28
Firmware 8.1 release notes added
Module limit raised to 16 · B16 support
Open the source page for any fact
Each result includes the supporting sentence, page number, document revision, and effective date.
Query corpus knowledge from your application
Full API reference{{ apiCode }}
{
"compatible": false,
"requirement": "firmware >= 7.4",
"sources": [
{ "doc": "compatibility-matrix-v8",
"page": 17 },
{ "doc": "service-bulletin-143",
"page": 2 }
]
}
{
"id": "MX900",
"family": "MX Series",
"revision": "C",
"lifecycle": "active",
"documents": 83,
"specifications": 184
}
{
"answer": "No.",
"detail": "B14 requires firmware
7.4 or newer.",
"confidence": "high",
"sources": [
"compatibility-matrix-v8#p17",
"service-bulletin-143#p2"
]
}
Schedule a demo
Bring a difficult document set.
Show us the files your current search, extraction, or RAG pipeline struggles with. We’ll walk through ingestion, structured knowledge, source evidence, and the API using your use case.