Find Companies
/find-companiesSearch companies by industry, size, revenue, funding, location and headcount growth, and get back matching accounts with their domains.
Overview
find_companies is the SyncGTM MCP tool that searches Crustdata and returns company profiles matching your firmographic filters — company name, primary domain, employee count range and country. It has no required parameter; you pass at least one filter drawn from identity, category, location, type, size, revenue, funding, LinkedIn following, age or headcount growth. This is the tool that builds an account list from scratch, as opposed to enrich_organization, which only works on domains you already have.
What it does
- 1
Takes no required input. You pass at least one filter, and stacking families — industry plus size plus country — is what turns a search into a workable list
- 2
The filters that carry most lists: industries and exclude_industries, employee_count_ranges, countries, company_types, min_revenue_usd, min_total_funding_usd, and last_funding_round_types paired with last_fundraise_after
- 3
Growth filters add timing: min_headcount_growth_percent_6m surfaces teams scaling fast, min_headcount_growth_absolute_12m surfaces large companies actually adding people
- 4
Returns matching company profiles — company_name, primary_domain, employee_count_range and countries — plus a next_cursor for the following page
- 5
Does not return people, emails or phone numbers. Chain find_people_within_company for named contacts and find_work_email for their addresses
- 6
Costs 0.3 credits per company returned. limit is 1–100 with a default of 25, so a default call is 7.5 credits and a full page of 100 is 30
How to use it
- 1
Download the skill file — it saves as find-companies.zip.
- 2
In Claude, open Settings → Capabilities → Skills and upload the .zip (or unzip the find-companies/ folder into .claude/skills/ for Claude Code).
- 3
Connect the SyncGTM MCP server from the setup guide — browser sign-in, no API key — then describe your account profile in plain language. The skill maps it to named parameters and shows you the mapping before it spends a credit.
- 4
Trigger it by typing /find-companies and pasting your ICP, for example: 'privately held Software Development companies in the United States, 51-200 employees, limit 50'.
- 5
The one setting worth tuning is limit. At 0.3 credits per company it is the cost dial, so set it explicitly and tell the skill how many pages to run rather than letting it page until the cursor runs out.
Use cases
Build a TAM list
Turn a written ICP into a list of domains you can actually work. One call replaces an afternoon of Sales Navigator exports and manual de-duplication.
Funding-triggered outbound
Pull companies whose last round was series_a or series_b and closed after a date you set. Reach them while the budget is new and the buying committee is still forming.
Headcount-growth prospecting
Find accounts that grew headcount 20%+ over the last six months. Fast-growing teams are the ones outgrowing the tooling they bought last year.
Investor portfolio mapping
Pass an investor name to list their portfolio companies, then run one pitch across a whole fund's book. Useful when a customer's investor also backs ten lookalikes.
GTM workflow examples
Monday morning funded-account blitz
- 1Call `check_credits` (free) to confirm the balance covers a 50-company search at 0.3 credits each — 15 credits before any enrichment.
- 2Call `find_companies` with industries `["Software Development"]`, countries `["United States"]`, employee_count_ranges `["51-200"]`, last_funding_round_types `["series_a", "series_b"]`, last_fundraise_after `"2026-02-01"`, limit 50 — produces up to 50 domains with size and country.
- 3Call `find_people_within_company` on each domain with job_title `["VP Sales", "Head of Revenue Operations"]` and max_profiles 2 — 1.5 credits per company, flat.
- 4Call `find_work_email` on each linkedin_url returned (1 credit each), then `verify_email` (0.3 each) and drop every row that comes back undeliverable.
Outcome: A sequencer-ready CSV of roughly 80 verified decision-makers across 50 US software companies that raised a Series A or B since February.
Growth-signal shortlist for the quarter
- 1Call `find_companies` with industries `["Financial Services"]`, countries `["United Kingdom"]`, min_headcount 200, min_headcount_growth_percent_6m 20, exclude_industries `["Staffing and Recruiting"]`, limit 40 — 12 credits, agencies already stripped out.
- 2Call `head_count_growth_rate` on the top 10 accounts (2 credits each) to separate a one-month hiring spike from twelve months of sustained expansion. It takes an `identifier` — a LinkedIn company id, LinkedIn URL or company name — not the domain find_companies hands back, so pass the company_name.
- 3Call `find_company_techstack` on those same 10 domains (1 credit each) and mark which ones already run a competing tool versus a legacy system you displace.
- 4Call `find_people_within_company` on the accounts that clear both checks, with job_title `["Chief Technology Officer", "VP Engineering"]` — 1.5 credits per company.
Outcome: A quarterly target list of up to 10 UK fintech accounts, each carrying a growth curve, a current stack and a named contact, built for under 60 credits.
Investor portfolio ABM tier
- 1Call `find_companies` with investors `["Sequoia Capital"]`, min_year_founded 2020, employee_count_ranges `["11-50", "51-200"]`, limit 60 — 18 credits for the fund's recent-vintage book.
- 2Call `enrich_organization` on the 15 closest-fit domains (2 credits each, 30 total) for industry, size, funding and technologies, and flag anything already sitting on a competitor's platform.
- 3Call `company_job_listings` on the domains that survive that cut (1.5 credits each) to find which are hiring the roles that signal budget for your category.
- 4Call `find_people_within_company` on the accounts with matching open roles, then `find_work_email` on each linkedin_url returned (1 credit each).
Outcome: An ABM tier drawn from 15 enriched Sequoia portfolio accounts, where every account left standing has a hiring-based reason to be contacted this week and a named, emailable owner.
Prompts
Paste any of these into Claude once the skill is installed and the SyncGTM MCP is connected.
Find privately held Software Development companies in the United States with 51-200 employees. Set limit to 50 and report the credits used.Find companies whose last funding round was series_a or series_b and closed after 2026-02-01, with employee_count_ranges 51-200 and 201-500. Limit 40, and show company name, domain, employee range and country.Find Financial Services companies in the United Kingdom with min_headcount 200 and min_headcount_growth_percent_6m of 20, excluding Staffing and Recruiting. Limit 30.Find companies founded after 2020 with min_total_funding_usd of 5000000, excluding Retail. Return 25, then page once more using the next_cursor and tell me the total credits spent.Find NASDAQ-listed companies with min_followers of 100000, limit 20, then run enrich_organization on the five largest domains.Best practices
Always pass at least one filter. There is no required parameter, so an unfiltered ask returns whatever the search defaults to and still bills 0.3 credits for every row it hands back.
Use the normalized size buckets: 1-10, 11-50, 51-200, 201-500, 501-1000, 1001-5000, 5001-10000, 10001+. The upstream Crustdata file also lists "myself only" and "2-10" — those do not match through the SyncGTM MCP.
industries and categories are exact strings from the Crustdata lists. "Software Development" and "Financial Services" match; "SaaS" and "Tech" are not in the list and quietly return nothing.
Dates must be YYYY-MM-DD, so convert "last six months" to last_fundraise_after: "2026-02-01" yourself. Growth windows are directional too — percent filters favor small companies, absolute filters favor large ones, _1m catches a spike and _12m catches sustained expansion.
Set limit explicitly and agree the page count before you start. Cost lives in paging, not in the call: three pages of 100 is 300 companies and 90 credits.
If you already hold the domains, do not search. enrich_organization (2 credits) or enrich_linkedin_page (0.5 credits) is the right tool — find_companies exists to produce domains, not to describe ones you have.
Filters and accepted values
Filters only match when the value is spelled exactly as the source list has it. Point the agent at these lists when you build a query.
- Industries
Flat array of 433 exact industry strings, e.g. "Software Development", "Financial Services". Used by industries and exclude_industries. "SaaS" and "Tech" are not in it.
- Crunchbase categories
Flat array for the categories parameter, e.g. "E-Commerce Platforms", "Food Delivery". Narrower than industries and worth pairing with it.
- Company types
Nine values for company_types: Partnership, Nonprofit, Privately Held, Self-Employed, Sole Proprietorship, Public Company, Educational Institution, Government Agency, null.
- Last funding round types
snake_case values for last_funding_round_types: pre_seed, seed, series_a through series_j, series_unknown, private_equity, debt_financing, post_ipo_equity, grant, angel, undisclosed and more. "Series A" with a space will not match.
- Employee count ranges
This file shows the upstream Crustdata buckets including "myself only" and "2-10". The SyncGTM MCP normalizes them — employee_count_ranges accepts 1-10, 11-50, 51-200, 201-500, 501-1000, 1001-5000, 5001-10000, 10001+.
- Largest headcount country
Array of {iso_alpha3, name} objects, e.g. {"iso_alpha3":"USA","name":"United States"}. Use for largest_headcount_countries when you care where the staff sit, not where the HQ is registered.