Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.parallel.ai/llms.txt

Use this file to discover all available pages before exploring further.

For AI agents: a documentation index is available at https://docs.parallel.ai/llms.txt. The full text of all docs is at https://docs.parallel.ai/llms-full.txt. You may also fetch any page as Markdown by appending .md to its URL or sending Accept: text/markdown.
Beta Notice: Entity Search is part of Parallel FindAll, which is currently in public beta. Endpoints and request/response formats are subject to change. We will provide 30 days notice before any breaking changes.
Entity Search finds people and companies on the web. It is a fast, synchronous API: you describe the people or companies you’re looking for in plain language and get back a structured set of matching results in seconds. Here, an entity is a real-world company or person. Entity Search is part of FindAll, and is designed for the parts of agentic systems that prioritize latency.
Entity Search is a fast, narrow lookup optimized for low latency. When you need comprehensive coverage — exhaustive list building, custom match conditions, enrichments, or per-field citations — use the FindAll API, its comprehensive, asynchronous sibling, instead.

​
When to Use It

Entity Search is built for real-time, human-in-the-loop, and latency-sensitive agentic workflows that need a candidate set of people or companies to evaluate, enrich, or pass into deeper research:
  • Interactive interfaces: Power a go-to-market or hiring interface, a chat experience, or an automation with a human in the loop, where results need to come back in seconds.
  • Starting sets for deeper work: Generate an initial set of people or companies to filter, enrich with additional fields, or hand off to FindAll or the Task API for comprehensive list building and research.
  • Cost-sensitive lookups: Run people-and-company search at a fraction of the cost of a full FindAll run when you don’t need exhaustive coverage or per-field citations.

​
Input and Output

You send a small set of inputs and receive a ranked, structured set of matching results. Input:
FieldTypeDescription
entity_typeenumThe kind of entity to search for: people or companies.
objectivestringNatural-language description of the people or companies you’re looking for.
match_limitintegerMaximum number of results to return. Between 5 and 1000 (inclusive); defaults to 100. May return fewer.
Output: an entity_set_id identifying the request, plus a ranked list of entities. Each entity includes its name, url, and description.
Entity Search does not support pagination — a request returns a single set of results, up to match_limit. To retrieve as many matches as possible, set match_limit to its maximum of 1,000.

​
Sample Request

curl -X POST "https://api.parallel.ai/v1beta/findall/entity-search" \
  -H "x-api-key: $PARALLEL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "entity_type": "companies",
    "objective": "AI startups that raised Series A in 2024",
    "match_limit": 100
  }'

​
Sample Response

{
  "entity_set_id": "entity_set_cad0a6d2dec046bd95ae900527d880e7",
  "entities": [
    {
      "name": "Figure AI",
      "url": "https://www.figure.ai",
      "description": "AI robotics company building general purpose humanoid robots"
    }
    // ... additional results omitted for brevity ...
  ]
}

​
How It Differs from FindAll

Entity Search and FindAll both turn natural-language criteria into entities, but they make different trade-offs. FindAll is the comprehensive, asynchronous sibling; Entity Search is the fast, narrow one.
Entity SearchFindAll
Input / output shapeA natural-language objective and an entity type in; a ranked list of name, url, and description outA structured schema with match conditions and optional enrichments in; verified, enriched records with reasoning out
CostMuch cheaper — priced per request (see Pricing)Higher — a fixed cost per run plus a per-match cost, and per-enrichment costs
SpeedMuch faster — synchronous, returning results in secondsAsynchronous, taking seconds to hours; poll, stream, or use webhooks
Fields it matches onA limited set: entity_type and the natural-language objective go in; only name, url, and description come back. No custom match conditions, enrichments, or citationsEvaluates arbitrary match_conditions, runs enrichments to extract additional fields, and returns per-field basis and citations

​
Scope and Compliance

Results draw on the public web and cover companies and people in a professional context — not consumer profiling, and not for use in employment, credit, or housing decisions.

​
Next Steps

  • FindAll Quickstart: Comprehensive, asynchronous entity discovery with verified, enriched, and cited results
  • Search API: Retrieve pages and excerpts across the whole web
  • Pricing: Detailed rate schedule
  • Rate Limits: Default quotas and how to request higher limits