Free US companies dataset sample: 10,000 companies, no sign-up
A free sample, no sign-up: 10,000 U.S. companies drawn from the free list, companies only, as CSV and Parquet, CC BY 4.0.
10,000 companies drawn at random from the companies of the free U.S. B2B leads dataset built in 2026-10 (1,748,139 companies with 11 to 500 employees and 9,521,154 people), one per state and per employee range first so every one is here.
Companies only: no person, no email, no phone. The street and postcode are left out, and so are self-owned companies and the companies abroad that the source calls U.S.: those in a city abroad filed under a U.S. state code (Perth, Western Australia as WA), and those with no U.S. address and no U.S. person (the next build leaves these out of the full file too).
Download it
- us_companies_sample_10k.csv (CSV, 7.6 MB)
- us_companies_sample_10k.parquet (Parquet, 3.8 MB)
- README.md (README, 1.6 kB)
Each is a plain link: no sign-up, no key, no link that expires. From a terminal, or straight into DuckDB:
curl -O https://datacircle.dev/samples/2026-10/us_companies_sample_10k.csv
curl -O https://datacircle.dev/samples/2026-10/us_companies_sample_10k.parquetimport duckdb
duckdb.sql("""
select NAME, URL, HQ_CITY, EMPLOYEE_COUNT_RANGE
from 'https://datacircle.dev/samples/2026-10/us_companies_sample_10k.parquet'
where HQ_STATE_CODE = 'TX'
""").show()The columns: 17, 10,000 rows
Each column with its fill rate in the sample, from its README. A field is filled only where we have it: nothing is guessed. The same columns in the full file, with their fill rates there: Dataset fields in the docs.
| Column | Type | Filled | What it is |
|---|---|---|---|
LINKEDIN_ID | String | 100% | Unique LinkedIn numeric ID for the company |
LINKEDIN_URL | String | 100% | Full LinkedIn company page URL (https://www.linkedin.com/company/...) |
NAME | String | 100% | Company name (acronyms normalized: LLC, PLLC, LLP, DBA, HVAC, USA) |
HEADLINE | String | 40% | Company tagline/headline from LinkedIn |
ABOUT | String | 64% | Company description/about section |
FOUNDED | UInt16 | 49% | Year the company was founded (validated: 1600–current year) |
LOGO_URL | String | 100% | URL to company logo image |
LINKEDIN_FOLLOWERS | UInt32 | 100% | Number of LinkedIn followers |
IS_CLAIMED_PAGE | Bool | 100% | Whether the LinkedIn page is claimed (always true - unclaimed pages are filtered out) |
EMPLOYEE_COUNT_RANGE | String | 100% | Employee count bucket. 10M+ dataset values: 11-50, 51-200, 201-500 |
HQ_CITY | String | 91% | Headquarters city |
HQ_STATE_CODE | String | 90% | Headquarters US state code (e.g. CA, NY) |
HQ_STATE_NAME | String | 90% | Headquarters US state name (e.g. California, New York). Empty string when unknown. |
LINKEDIN_INDUSTRY | String | 90% | LinkedIn industry classification |
TYPE | String | 58% | Entity type: Privately Held, Public Company, Partnership, Self-Owned |
URL | String | 80% | Company website URL |
UPDATED_AT | DateTime | 100% | When the record was last refreshed |
License
CC BY 4.0: reuse it freely, crediting datacircle.dev. To cite it: Datacircle, “10,000 U.S. companies: a sample of Datacircle's free 10M+ U.S. B2B leads dataset (2026-10)”, https://datacircle.dev/us-companies-dataset-sample.
The whole dataset
Free: 10M+ U.S. B2B leads, as a flat file. Download it at datacircle.dev. Every column of it, with the people: the free US B2B leads dataset. The companies by state, industry and size: the free list of US companies.
Questions
Do I need to sign up to download the sample?
No. The CSV, the Parquet and the README are plain links on this page: no sign-up, no key, no link that expires. A free sample, no sign-up: 10,000 U.S. companies drawn from the free list, companies only, as CSV and Parquet, CC BY 4.0.
Can I use it in a commercial project?
Yes. It's under CC BY 4.0: reuse it freely, commercial use included, crediting datacircle.dev.
How is it different from the full dataset?
The sample is 10,000 companies and nothing else. Free: 10M+ U.S. B2B leads, as a flat file. Download it at datacircle.dev. The full file holds the people too, with work emails and mobile phones where we have them, and the companies' street and postcode.
How were the companies picked?
One company per state and per employee range first, so every one is there, then the rest at random with a fixed seed, so it is spread like the full file.
Sign up at datacircle.dev with your work email: a $5 credit, that's 4,000 LinkedIn profiles at $1.25 per 1,000. Free: 10M+ U.S. B2B leads, as a flat file. Download it at datacircle.dev.
Sign up