Skip to main content
Your assistant picks the right Swarm tools based on your question — you never call them directly. This page shows how to prompt for the best results, then lists every tool the assistant can use.

​
Tools reference

The Swarm MCP exposes nine tools, in two groups.

​
Action tools

These search The Swarm and return results. Three of them are billable — see Credits and cost for the full rules.

​
find_profiles

Search The Swarm’s people database. Returns full profiles — current and past jobs, education, skills, locations, emails, and social links. When the assistant uses it. Whenever you ask to find, filter, list, or summarize people. Cost. 1 credit per successful search (empty results are free).

​
find_companies

Search The Swarm’s company database. Returns full company records — industry, size, locations, funding rounds, business status, and social links. When the assistant uses it. Whenever you ask about companies as a group. Cost. 1 credit per successful search (empty results are free).

​
relationships

Answers “who do we know at…?” — finds people in your team’s network, showing which teammates know them and how strong the connection is. When the assistant uses it. Any time you ask about warm paths, introductions, or connections through your team. Cost. Free.

​
partner_network_mapper

Search inside another company’s network by their website domain. If the company isn’t mapped yet, this starts a background mapping task and returns a task ID.
Requires partner access.
When the assistant uses it. When you ask to “map” a company or partner and then search inside their network. Cost. 1 credit per mapping that returns matches (empty results are free).

​
Helper tools

Setup and lookup tools the assistant calls before or alongside the action tools — they don’t return end-user results on their own.

​
network_mapping_status

Check the progress of a network-mapping task started by partner_network_mapper.
Requires partner access.

​
profile_dictionary

Look up the exact spelling of people filter values (job title, seniority, industry, connection strength, pipeline stage, etc.) before filtering a search.

​
company_dictionary

Look up the exact spelling of company filter values (industry, location, size, investor type, etc.) before filtering a search.

​
get_profile_mapping

Returns the list of searchable fields for people. Helper step for building a people query.

​
get_company_mapping

Returns the list of searchable fields for companies. Helper step for building a company query.

​
Notes for developers

  • Action tools accept OpenSearch DSL queries. Your assistant constructs these for you based on your natural-language request.
  • Partner-access tools (partner_network_mapper, network_mapping_status) are only available on plans with partner access.
  • Dictionary and mapping tools return values used to build accurate filtered queries — the assistant calls them automatically as needed.

​
Prompting tips

  • Be specific. “Series A fintech founders in Berlin” beats “founders in Germany.”
  • Layer your filters. Add role, seniority, industry, location, or company one at a time and refine from there.
  • Ask for the network angle. After finding people or companies, ask “who on our team is connected to them?” or “who can introduce me?”
  • Use exact place names. Use a location’s canonical local name (for example, Warszawa rather than Warsaw) for the most reliable matches.
  • Empty result? That usually means the filter was too narrow or a value wasn’t recognized — try rephrasing or broadening.
  • Follow up. Narrow the location, change the seniority, or ask the assistant to draft an intro message.

​
Primary use cases

​
Sales prospecting and warm intros

Find decision-makers at your target accounts, then discover who on your team can make the introduction.
“Find heads of RevOps at Series B SaaS companies in the US that raised in the last 12 months. Then check which ones anyone on our team is connected to.”
What you’ll get back: A shortlist of matching people, followed by a table of who in your network knows them and how strong the connection is.

​
Recruiting shortlists

Build candidate lists filtered by role, seniority, skills, and location.
“Find senior ML engineers in Berlin or Amsterdam with experience at scale-ups.”
What you’ll get back: A ranked list of profiles with role history, skills, and location. Ask for a summary comparison of the top candidates.

​
Partnership and account mapping

Map a target organization’s people and find the strongest connection paths.
“Map the network at acme.com, then show me the most senior people in Product and Engineering, and who on our team knows them best.”
What you’ll get back: Confirmation that mapping is running (large maps happen in the background), then results grouped by function and seniority, with connection strength.

​
Competitive and market research

Search companies by industry, size, funding, and geography.
“List climate-tech companies in the Nordics with 50–200 employees that raised funding in the last year.”
What you’ll get back: A list of companies with industry, headcount, location, and latest funding, ready for follow-up questions like “which of these do we have connections into?”

​
Example prompts

Find people
“Find VP-level product leaders at Series B SaaS companies in London and summarize the top 10.”
Find companies
“List climate-tech companies in the Nordics with 50–200 employees that raised funding in the last year.”
Discover warm paths
“Who on our team is connected to someone at Stripe, and how strong are those connections?”
Map a partner’s network
“Map the network at acme.com, then show me the most senior people once it’s ready.”
Draft an intro
“Based on the strongest connection you just found, draft a short intro request I could send to my teammate.”