Concepts
The objects you work with through the API, and how they relate.
Agent
Collection for one source — a court portal, a registry, a licensing board — engineered, deployed, and maintained by NexqData. You run agents; you don't write them. Each agent has a version, a set of input parameters, and two ids:
| Field | Type | Used for |
|---|---|---|
id | GUID | Reading the agent, its runs, and its jobs. |
externalAgentId | string | Starting a job (or a single run). |
You can only see and run agents assigned to you (or your company).
Job
One start of an agent, and the recommended way to start one — it's how the NexqData platform starts agents, from the dashboard to groups, pipelines, and schedules. Starting a job returns immediately with its job id (agentJobHistoryId); the work finishes in the background.
A job creates one or more runs, depending on how the agent is built — for example, an agent that splits a large search across several sessions. The job reports their combined status, and its export combines their data. See Jobs.
Run
One execution of an agent. Every run a job creates is listed under it, and each has its own run id (agentRunHistoryId), status, and data. A run is active until it reaches a terminal state, then succeeded or failed. See Run status & history.
Agent group
Agents that cover the same kind of record — every county, every state board — run together with one set of inputs. Each member runs independently and reports its own status. See Agent groups and Live Search.
Pipeline
Agents in a fixed order. Each stage starts only after the previous one succeeds; a failure stops the chain. Every stage's output downloads together. See Pipelines.
Dataset
A run's structured output, stored by NexqData. Datasets are private to the user who collected them by default, and can be shared with the user's company. You can preview, query, and download them. See Datasets.
Schedule
A recurring (or one-time) trigger for an agent, group, or pipeline. See Schedules.