Lookalike Companies

Lookalike Companies is a Claude skill that takes one seed domain and returns companies that look like it. It reads the seed's profile from LeadOcean, then searches the company database on the same industry, size and country.

What it does

You give Claude a domain, for example acme.com. The skill pulls that company's industry, headcount range, headquarters and funding. It then searches for other companies with the same profile and drops the seed from the results.

The match is firmographic. It uses filters, not a similarity score, so every row is there for a reason you can read. Add match_funding: yes to also match the last funding round.

It runs on the LeadOcean MCP server, so there is no API key to paste. You sign in once with OAuth.

How it works

The skill is a SKILL.md file. It tells Claude which tools to call and in what order.

1. get_company reads the seed: industry, employee range, headquarters country, funding. 1 record.
2. Claude builds the filters from that profile. No tool call.
3. list_enum_values resolves the exact industry string if a value is rejected. Free.
4. search_companies returns up to 25 companies on those filters, minus the seed. 1 record per company returned.
5. If fewer than 10 come back, Claude drops one filter at a time and searches again. It tells you which filter it dropped.
6. count_leads sizes the mailable C-Team and VP people at the top 5 companies. Free.

Claude ranks the list so companies matching on size, country and funding come first. People are not pulled unless you ask.

How to use it

1. Download the zip: Download lookalike-companies.zip.
2. Unzip it and add the lookalike-companies folder as a skill in Claude.
3. Connect LeadOcean at https://api.leadocean.io/mcp and sign in.
4. Ask: "Find lookalike companies for acme.com". Add options if you want: "in DE, 40 results".

You need a LeadOcean account. The free plan gives you 1,000 records, one-off, no card.

Sample output

The sizing step is free, so we ran it. On 1 October 2026 we looked up the exact industry name with list_enum_values (Software Development is one of seven software values). Then we ran count_leads on a lookalike profile: a Series B software company in the US with 201-500 employees.

The count covers mailable people at matching companies, not the companies themselves. The free count does not take the funding filter, so the skill applies fundingType at search time and the list gets shorter than this number. The company list needs search_companies, which spends records, so no real list is pasted here. This is the shape of the reply, with placeholder values.

Output format (placeholder values)

Lookalikes for Acme (acme.com)
As of 1 October 2026

Seed profile
- Industry: Software Development
- Size: 201-500 employees | HQ: US | Founded: 2015
- Funding: Series B

Filters used
industry, employeeRange, hqCountry

Lookalike companies
| Company | Domain | Industry | Size | HQ | Funding |
| Example Co | example.com | Software Development | 201-500 | Austin, US | Series B |

Next step
Mailable C-Team and VP people at the top 5: 120. Ask if the user wants them pulled.

Records spent: 26

The skill replies in this shape and nothing else. The title line is a Markdown H1 and each section title is an H2. Any field it cannot find reads "not on file".

Records per run

The free plan's 1,000 records cover about 38 default runs. A widening search adds its own rows, so a run can cost more. On Pro ($499 a month) there is no per-record price. Fair use applies, as described on the pricing page.

Related skills

• Account Researcher: a brief and buying committee for one company.
• Funding Trigger: companies that raised recently.
• Contact Finder: the people at a company you picked.
• All Claude skills

Background reading: the best lookalike company search tools, find company lookalikes from HubSpot and lookalike audiences.

FAQ