Skip to main content
This glossary covers the terms you’ll see throughout the dashboard, the API, and the rest of these docs. Entries are alphabetical; click any term to expand its definition.

​
A

A pipeline node that processes data between a trigger and an output. Actions include parse, extract, classify, route, merge, review, transform, validation, and HTTP action.
The state that lets a pipeline accept new documents and execute runs. Activation requires a valid configuration with at least one trigger. See pipelines.
A monitoring rule that fires a notification when a metric crosses a threshold (for example, a failure rate above 10% or a credit balance below 500). See alert rules.
A credential that authenticates programmatic requests to the Ingestly API. Keys use the ing_ prefix and are passed in the X-Api-Key header. See API keys.

​
C

A webhook notification Ingestly sends to an external URL when a run completes. Configured with the callback output node.
A run created to process a single chunk of a document split by a parent run. The parent waits in the Awaiting Children status until every child run finishes. See runs.
An OAuth integration that lets pipeline nodes (like HTTP action) call an external service without exposing credentials. See connectors.
The billing unit consumed per pipeline run. Each node type has its own credit cost; see credits for the full breakdown.

​
D

A file uploaded to a pipeline for processing. Ingestly supports PDF, JPEG, PNG, TIFF, and HEIC. See documents.
Splitting a multi-page document into separate chunks that each run through the pipeline as a child run. See runs.

​
F

Configurable behavior on a validation node when validation errors occur. Fail stops the run; Warn lets the run continue with warnings attached.
A trigger that receives forwarded documents and reference samples for AI similarity classification. See forward trigger.

​
N

A single step in a pipeline. Nodes are grouped into three categories: triggers, actions, and outputs.

​
O

The top-level account in Ingestly. Each organization holds workspaces, users, API keys, billing, and a credit balance.
A pipeline node that delivers final results out of Ingestly. Outputs include callback, email output, and forward output.

​
P

The named workflow you configure for document ingestion. A pipeline holds its configuration (a directed graph of nodes and edges), its documents, and its run history. Surfaced consistently across the UI, the API, and runs. See pipelines.
The visual canvas where you build a pipeline by dragging in nodes and connecting them. See pipeline editor.
A workspace-scoped grouping that lets you organize related pipelines by name and color on the Pipelines page. See managing pipelines.

​
R

The number of days Ingestly keeps processed documents and run data for a pipeline. Configured per pipeline.
The list of paused runs waiting for a human to inspect and approve. See review queue.
A pause point in a run created by a review node. A reviewer can edit and approve the data before the run continues.
One execution of a pipeline, triggered when a document is submitted to it. See runs.

​
S

The JSON Schema (Draft-07) definition that tells the extract node which fields to pull from a document. See schema design.
The result of a single node within a run. Captures input, output, status, timing, and any errors for that step. See step executions.

​
T

The first node in a pipeline. Defines how documents enter the pipeline: upload, email, webhook, or forward.

​
U

A field whose extraction confidence falls below an internal threshold when Field confidence is enabled on the extract node. Pair with a review node configured to Pause mode = Unverified Fields to send uncertain runs to human review.

​
W

The HMAC-SHA256 secret Ingestly uses to sign callback requests so your application can verify they came from Ingestly. See webhook signing keys.
An isolated container for pipelines, documents, and runs inside an organization. Pipelines in one workspace are not visible from another. See workspaces.