Skip to content
nexqdata
Live search

Turn difficult websites into APIs.

Search courts, registries, portals, and other complex sources from your own application — and get structured records back.

Watch a search

One request, three sources searched in parallel, one response. Illustrative — timings and records are examples.

Illustrative live search: your application sends one search request to NexqData, which searches three sources in parallel — a county court portal, a state business registry, and a licensing board that times out once and answers after a retry. Five records come back from the three sources, and NexqData returns them to your application in 7.1s with a 200 OK response.

How live search works

Request. Execute. Structure. Respond.

Four stages between your application asking and your application knowing.

  1. 01

    Request

    POST /search

    Your application sends the inputs it has — a name, a date of birth, a case number — and which agent or group of sources to search.

  2. 02

    Execute

    3 sources · in parallel

    NexqData runs the agent engineered for each source. Sources run independently, so one slow portal never holds up the rest.

  3. 03

    Structure

    records[] per source

    Each source returns records in the fields agreed when its agent was scoped — not pages of HTML to parse.

  4. 04

    Respond

    200 OK · status per source

    Your application gets the results through the API, with a status for every source, so you know exactly which ones answered.

Multi-source

One request. Many sources.

Name a group instead of a single agent, and the same request searches every source in it.

  • Independent. Each source runs on its own; a slow one never holds up the rest.
  • Accountable. Every source reports its own status, so a gap is visible, not silent.
  • Extendable. Add a source to the group, and every search includes it.

For developers

One request in. Structured records out.

~/app — zsh — 120×32
# Search a source — or a whole group — from your application$ time curl -s -X POST "$NEXQDATA_API/search" \>   -H "X-Api-Key: $NEXQDATA_API_KEY" \>   -H "Content-Type: application/json" \>   -d '{"agent": "court-records-tx", "input": {"firstName": "John", "lastName": "Doe", "dob": "1990-01-01"}}' \>   | jq{  "status": "completed",  "sources": [    {      "source": "court-records-tx",      "status": "completed",      "records": [        {          "caseNumber": "2026-0418823",          "court": "District Court 190",          "filed": "2026-09-22",          "caseStatus": "Open",          "matchedOn": ["firstName", "lastName", "dob"]        }      ]    }  ]} real    0m7.104suser    0m0.021ssys     0m0.012s$ 

Illustrative. Endpoint names, input fields, and the shape of each source’s records are agreed with you when we scope your agents.

Scheduled or live

Search when you need the answer.

Both run the same engineered agents. The difference is how old the answer is when someone asks.

Scheduled collection

Periodic
Data age when asked12h 37m

Agents run on a schedule and refresh a dataset. You query your copy, which is as fresh as the last run.

Best for: Feeds, monitoring, full datasets

Scheduling on the platform

Live search

On demand
Data age when asked7.1s

Agents run when you ask. The answer comes from the source at that moment, for the inputs you sent.

Best for: Screening, lookups, decisions made now

See the request

Built for production

Fast answers, from agents built to last.

Engineered

Every source gets its own agent, built around its forms, sessions, and quirks.

Learn more →

Multi-source

Group agents that cover the same kind of record and search them all at once.

Learn more →

Monitored

Failed or unusual searches are flagged, and agents are updated when a source changes.

Learn more →

Access

Per-user API keys that can be scoped to specific permissions and revoked at any time.

Learn more →

Which sources do you need to search?

Tell us the sources and the inputs you have. We’ll scope the agents and the shape of the results with you.