Market Map via Similar-Companies Crawl
Crawl one seed's similar-companies graph on LinkedIn, then cross-fill the gaps with AI competitor lists and funding news into a full, sized market map.
https://apidirect.io/mcp?token=YOUR_API_KEY
Use the market-map-similar-companies skill: Build me a map of {seed_company}'s competitive landscape by expanding similar companies two hops out.
market-map-similar-companies) makes your agent run this exact playbook instead of improvising its own searches.
In clients that support MCP prompts (Claude Desktop, Claude Code, Cursor) it also appears as a prompt named
market-map-similar-companies, and any agent can call
get_skill(skill_id="market-map-similar-companies") directly.
What it does
Walks LinkedIn's similar_companies edges breadth-first from one seed, enriching each with size, founding year and specialities. Optionally cross-fills the graph's blind spots with an AI competitor list (web) and a category funding-news scan (news), deduping everything into one sized landscape — with the user choosing which platforms run.
Who it's for: Strategy, corp-dev, competitive-intel and founders mapping a space.
Inputs
| Input | Required | Description | Example |
|---|---|---|---|
seed_company |
Yes | A representative company to start the crawl from. | Ramp |
depth |
No | How many hops to expand the LinkedIn similar_companies graph (2 is usually plenty). | 2 |
platforms |
No | Comma-separated list of platforms to run — options: linkedin, web, news. Omit to run all of them; name specific platforms to limit the run. | linkedin, web, news |
region |
No | Optional country focus for the AI competitor list and news scan (maps to the country param). Leave blank for global. | us |
news_window |
No | How far back to scan funding/entrant news (maps to time_published: 1h, 1d, 7d, 1y, anytime). | 1y |
How your agent runs it
-
1
search_linkedin_companies(query="{seed_company}", page=1)Resolve the seed to a company page URL.
-
2
linkedin_company_details(url=<seed_company_url>)Read similar_companies[], employee count, founded_year and specialities. Capture the seed's category/specialities to steer the later web and news expansion. Queue each similar company you haven't seen.
-
3
linkedin_company_details(url=<next_similar_company_url>)Repeat breadth-first up to {depth} hops, deduping by company. Stop when no new companies appear.
-
4
google_ai_mode(prompt="List companies that compete with or are similar to {seed_company} in <seed_category>")Only run if {platforms} includes web: surface private, regional and non-LinkedIn-active players the similar_companies graph misses. For each new company name, feed it back through steps 1-3 to resolve its URL and enrich it (tag source = AI list).
-
5
search_news(query="<seed_category> startup OR {seed_company} competitor funding", time_published="1y", limit=50)Only run if {platforms} includes news: catch newly funded or emerging entrants in the category. Add any new company to the map flagged as a new entrant and enrich via steps 1-3 (tag source = news).
Tools used
The full playbook
This is exactly what the MCP returns to your agent (via the market-map-similar-companies prompt or get_skill tool), with your inputs filled in.
SKILL: Market Map via Similar-Companies Crawl
Walks LinkedIn's similar_companies edges breadth-first from one seed, enriching each with size, founding year and specialities. Optionally cross-fills the graph's blind spots with an AI competitor list (web) and a category funding-news scan (news), deduping everything into one sized landscape — with the user choosing which platforms run.
You are running this skill on API Direct via its MCP tools. Execute the steps below yourself by calling the named tools in order — values in <angle brackets> come from a previous step. Then deliver the result described at the end.
INPUTS:
- seed_company (required): <seed_company — ASK THE USER>
A representative company to start the crawl from.
- depth (optional): (optional — e.g. 2)
How many hops to expand the LinkedIn similar_companies graph (2 is usually plenty).
- platforms (optional): (optional — e.g. linkedin, web, news)
Comma-separated list of platforms to run — options: linkedin, web, news. Omit to run all of them; name specific platforms to limit the run.
- region (optional): (optional — e.g. us)
Optional country focus for the AI competitor list and news scan (maps to the country param). Leave blank for global.
- news_window (optional): 1y (default — override if you like)
How far back to scan funding/entrant news (maps to time_published: 1h, 1d, 7d, 1y, anytime).
PLATFORM SELECTION: some steps are gated with "Only run if {platforms} includes X". If the user supplied a `platforms` value, run only the steps whose platform is listed. If the user did NOT supply `platforms`, run ALL steps — every platform the skill supports.
STEPS:
1. Tool `search_linkedin_companies` — search_linkedin_companies(query="{seed_company}", page=1)
Resolve the seed to a company page URL.
2. Tool `linkedin_company_details` — linkedin_company_details(url=<seed_company_url>)
Read similar_companies[], employee count, founded_year and specialities. Capture the seed's category/specialities to steer the later web and news expansion. Queue each similar company you haven't seen.
3. Tool `linkedin_company_details` — linkedin_company_details(url=<next_similar_company_url>)
Repeat breadth-first up to {depth} hops, deduping by company. Stop when no new companies appear.
4. Tool `google_ai_mode` — google_ai_mode(prompt="List companies that compete with or are similar to {seed_company} in <seed_category>")
Only run if {platforms} includes web: surface private, regional and non-LinkedIn-active players the similar_companies graph misses. For each new company name, feed it back through steps 1-3 to resolve its URL and enrich it (tag source = AI list).
5. Tool `search_news` — search_news(query="<seed_category> startup OR {seed_company} competitor funding", time_published="1y", limit=50)
Only run if {platforms} includes news: catch newly funded or emerging entrants in the category. Add any new company to the map flagged as a new entrant and enrich via steps 1-3 (tag source = news).
DELIVER: A deduped market table — company, size, founding year, specialities, and source (LinkedIn graph / AI list / news) — grouped into clusters, with newly funded or non-LinkedIn entrants flagged.
Note: each underlying tool call is billed at its normal endpoint price; get_sentiment adds a small per-page surcharge. Page through results as needed but stop once you have enough to deliver the outcome.