FindAll API
# Find the best data for your database, contact list, company directory, press list, talent list, competitive set, financial report
## Discover and catalogue any set of entities from a single natural language web query
## Stop buying stale data
Create dynamic datasets on demand
FindAll discovers and structures any set of entities: companies, people, properties, or events, from a single natural language query
Build lead lists, source deals, map markets, and create custom datasets in minutes instead of weeks
## Effortlessly extend your datasets
Use the FindAll API to discover entities based on your specified criteria
Then use the Task API to enrich your data with new columns that expand the depth of your database
Match compute budget with query difficulty
Not every dataset requires the same depth. FindAll offers tiered Processors and Generators[Processors and Generators](https://docs.parallel.ai/findall-api/core-concepts/findall-generator-pricing) that let you match AI compute to research complexity, optimizing for cost, speed, and thoroughness.
The best recall at every price point. From simple lookups to analyst-grade research.
CPM: USD per 1000 requests. Cost is shown on a Log scale.
### Benchmark
This benchmark, created by Parallel, contains 40 complex multi-criteria queries covering public companies, startups, SMBs, specialized entities, and people (e.g., executives, researchers, professionals).
### Methodology
To measure recall we take the number of correct matches / total entities in the ground truth dataset. The ground truth dataset is created by taking the union of all correct matches across the competitor set. Cost is calculated as the average cost to find 1000 correct matches.
### Testing dates
Nov 13th-17th, 2025
## Match compute budget with query difficulty
Not every dataset requires the same depth. FindAll offers tiered Processors and Generators[Processors and Generators](https://docs.parallel.ai/findall-api/core-concepts/findall-generator-pricing) that let you match AI compute to research complexity, optimizing for cost, speed, and thoroughness.
The best recall at every price point. From simple lookups to analyst-grade research.
### Parallel-FindAll
| Series | Model | Cost (CPM) | Recall (%) | | -------- | ----------------------- | ---------- | ---------- | | Parallel | FindAll Base | 60 | 30.3 | | Parallel | FindAll Core | 230 | 52.5 | | Parallel | FindAll Pro | 1430 | 61.3 | | Others | OpenAI Deep Research | 250 | 21 | | Others | Anthropic Deep Research | 1000 | 15.3 | | Others | Exa | 110 | 19.2 |
CPM: USD per 1000 requests. Cost is shown on a Log scale.
### Benchmark
This benchmark, created by Parallel, contains 40 complex multi-criteria queries covering public companies, startups, SMBs, specialized entities, and people (e.g., executives, researchers, professionals).
### Methodology
To measure recall we take the number of correct matches / total entities in the ground truth dataset. The ground truth dataset is created by taking the union of all correct matches across the competitor set. Cost is calculated as the average cost to find 1000 correct matches.
### Testing dates
Nov 13th-17th, 2025
Apply FindAll to a broad range of business needs
FindAll powers any workflow that starts with "Find all the..."
Find all companies that are GDPR-compliant and operate in more than three countries
Find all companies in California which are SOC-2 compliant
Find all residential roofing companies in Palo Alto
Find all companies that increased their market cap by more than 20% in the past year
Find all companies founded after 2020, and whose founders are former Google researchers
Find all dental practices in California that are currently accepting new patients
Find all companies that hired a new CFO in the past 6 months and are based in USA
1. Build hyper-targeted lead lists
Go beyond firmographics. Find companies based on signals that actually matter—recent leadership changes, technology adoption, expansion announcements, or any criteria specific to your ICP.
2. Find acquisition or investment opportunities
Identify acquisition targets, investment opportunities, and market entrants based on precise criteria—funding stage, employee count, geography, partnerships, or competitive positioning.
3. Map markets in real time
Track new entrants, monitor competitor movements, and build comprehensive market landscapes. Update your maps continuously as the market evolves.
4. Find suppliers that meet your exact requirements
Discover manufacturers, service providers, and partners filtered by location, capacity, certifications, or any other criteria critical to your operations.
5. Monitor regulatory and legal landscapes
Track litigation, enforcement actions, regulatory filings, and compliance-relevant events across jurisdictions and industries.
Build hyper-targeted lead lists
Go beyond firmographics. Find companies based on signals that actually matter—recent leadership changes, technology adoption, expansion announcements, or any criteria specific to your ICP.
# Get started
with just a few lines of code
from parallel import Parallel
client = Parallel(api_key="YOUR_API_KEY")
findall_run = client.beta.findall.create(
objective="FindAll portfolio companies of Khosla Ventures founded after 2020",
entity_type="companies",
match_conditions=[
{
"name": "khosla_ventures_portfolio_check",
"description": "Company must be a portfolio company of Khosla Ventures."
},
{
"name": "founded_after_2020_check",
"description": "Company must have been founded after 2020."
}
],
generator="core",
match_limit=5
)
## FAQ
FindAll pricing combines a fixed cost per run with a per-match cost, both determined by your chosen _Generator._ If you add enrichments to extract additional data from matches, those add per-match costs based on the Task API processor tier you select.
Learn more about Generators here: https://docs.parallel.ai/findall-api/core-concepts/findall-generator-pricing[https://docs.parallel.ai/findall-api/core-concepts/findall-generator-pricing](https://docs.parallel.ai/findall-api/core-concepts/findall-generator-pricing)
FindAll is an API that discovers and returns structured datasets of entities—companies, people, properties, events, and more—from across the web based on natural language queries with multi-criteria filters. It transforms the web into an on-demand database tailored to your specific requirements.
FindAll can find virtually any type of entity that exists on the web: companies (public, private, startups, SMBs), people (executives, researchers, professionals), real estate properties, events, products, legal cases, academic papers, and more. If information about it exists on the web, FindAll can discover and structure it.
Traditional data providers sell pre-built, static datasets that become stale quickly and may not match your specific criteria. FindAll generates datasets on demand based on your exact requirements, sourced from the live web. Your criteria can be as specific or complex as needed, and you can regenerate datasets whenever you need fresh data.
Web scraping requires you to identify target sites, build and maintain extraction code, handle anti-bot measures, and structure the output yourself. FindAll handles all of this automatically—you simply describe what you're looking for in natural language, and receive structured, verified results. No infrastructure to maintain, no code to update when sites change.
Yes, Parallel supports “multi-hop” reasoning. This means FindAll can evaluate match criteria that require synthesizing information from multiple web sources. For example, to match "healthcare startups with Series A funding and FDA-approved products," FindAll might need to verify the company's industry from one source, funding status from another, and FDA approval from a third. This enables complex, compound queries that simple keyword search cannot handle.
In benchmarks against OpenAI Deep Research, Anthropic Deep Research, and Exa, FindAll Pro achieved 61% recall—approximately 3× higher than alternatives. Recall measures the proportion of all correct matches that FindAll successfully identifies. Every result includes citations, source excerpts, and confidence scores so you can verify accuracy for your use case.
Every entity FindAll returns includes the sources used to verify it matches your criteria, excerpts from those sources, the reasoning behind the match decision, and a confidence score indicating how certain the match is. This enables human-in-the-loop verification workflows and provides full provenance for compliance-sensitive applications.
Yes. FindAll includes an enrichment capability that extracts additional structured fields from matched entities—revenue, employee count, technology stack, recent news, strategic initiatives, or any other attribute you specify. Enrichments are powered by the Parallel Task API and can be made to run automatically on every match.
FindAll offers multiple tiers optimized for different use cases. Base is the most cost-effective option for broad queries with many expected matches. Core balances cost and thoroughness for moderately specific queries. Pro provides the most comprehensive search for highly specific queries with rare or hard-to-find matches. See our documentation for detailed guidance on choosing the right tier.
You can test FindAll immediately in our Playground without writing any code. When you're ready to integrate, our documentation provides complete API references, code examples, and best practices for production use.
Yes. FindAll offers a Preview mode that runs your query against approximately 10 candidates so you can validate your approach and estimate match volume before committing to a full run.