for ai agents · skills · mcp
stop scraping.
your agent has the data.
structured datasets your agent can filter and pull — agencies, product & business directories, newsletters — with traffic, DR, subscribers, and ratings attached. one skill, one api key, many datasets.
$ npx skills add nostrband/servicegraphfree tier — 2,000 credits on signup, no card.
Available for 19 agents
what it looks like
one prompt, one filter, real rows — the same interface across every dataset. agency shortlists, launch directories, newsletters to sponsor.
user> Find me three boutique IP law firms in California that handle patent prosecution for hardware startups.
[skill: find-service-providers]
GET /v1/datasets/pro_services/search?filter=industry:legal+state:CA+service_provided:patent-law → 37 rows in pool POST /v1/datasets/pro_services/unlocks for top picks
- • Knobbe Martens— knobbe.com — (949) 760-0404 ★★★★★ 4.7 (Clutch · 142)
- • Carr & Ferrell— carrferrell.com — (650) 812-3400 ★★★★★ 4.5 (G2 · 31)
- • Hickman Becker Bingham— hickmanlaw.com — (650) 314-0700 ★★★★★ 4.6 (Clutch + Avvo · 87)
user> Use ServiceGraph to find product directories with DR 80+ where I can submit a dev-tools SaaS.
[skill: servicegraph]
GET /v1/datasets/product_directory/search?filter=dr>=80+keyword:submit → 60+ rows in pool POST /v1/datasets/product_directory/unlocks for top picks
- • SourceForge — sourceforge.net — DR 92 software
- • Product Hunt — producthunt.com — DR 91 launch
- • G2 — g2.com — DR 91 reviews
- • Startup Fame — startupfa.me — DR 82 launch
one dsl, many datasets
agencies are one dataset of many. the branded servicegraph skill discovers every one at runtime — same filter DSL, same credit balance.
discover datasets, inspect schemas, validate filters, and search brief rows — identity plus headline metrics. browse as much as you like.
unlock a row to reveal contacts, full metrics, and sources. 10 credits per row on current datasets; unlocks cached, re-unlock within the window is free.
what agents do with it
find product directories and launch sites your audience actually visits — filter by DR and traffic, skip the dead ones.
newsletters to sponsor and agencies to pitch, with contacts, sortable by niche, subscribers, services, and ratings.
shortlist agencies by services, size, location, and third-party ratings — signals beyond a google search.
pass a list of domains, get back classified industry, metrics, and contacts in one structured response. no search step required.
sample prompts
click any prompt for the three-step run-through. naming “ServiceGraph” routes to the branded skill (any dataset); a bare ask routes to the matching agency shortcut.
faq
Yes. Every call from a skill, MCP server, or your own script authenticates with a key. Create one at /profile/api-keys after signing in and set it as SERVICEGRAPH_API_KEY. 2,000 free credits on signup, no card.
Whichever your harness supports; all hit the same backend. The branded servicegraph skill works against every dataset, the find-* skills are agency shortcuts, the MCP server at mcp.servicegraph.co handles OAuth in-sandbox, or call the REST API directly.
19+ — Claude Code, Codex, Cursor, GitHub Copilot, Gemini, Cline, Goose, Windsurf, and any harness that supports the Agent Skills format. No skills support? Call the API directly.
Yes. The branded servicegraphskill discovers every dataset at runtime — agencies, product & business directories, newsletters, and more as they ship. The find-* skills cover agencies today; per-dataset shortcuts land as we build them.
Name, url, the headline metrics for that dataset, plus contacts where we have them. Search and brief rows are free across every dataset; only detail unlocks spend credits.
10 credits to unlock 1 row on current datasets (≈ $0.10, $0.08 at volume). Signup credits ≈ 200 free rows. Per-row price may vary on future datasets — your agent reads it from the API, never assumes it.
30 days from unlock. Re-unlocking within that window is free.
Built from public web data and refreshed regularly; cadence varies by dataset. Classification is re-checked on update, stale rows are dropped.
Yes — CSV on the web, JSON via the API.
GET /v1/datasets/:id/:apex returns a 404 (uncharged) so your agent can branch. Need something indexed? Email artur@servicegraph.co and we can prioritize.
plug it in. start with 2,000 free credits.
that’s 200 rows free. no card. no subscription.