ZoomInfo MCP Server

MCP server for the ZoomInfo ZI API (the new developer.zoominfo.com platform) — company/contact enrichment, search, intent, scoops, news, taxonomy lookups, and Copilot insights.

Read-only. Every tool maps to a read/discovery endpoint; no ZoomInfo records are created or modified.

Authentication

OAuth2 client credentials (server-to-server). The server mints a bearer token from your DevPortal app's client_id / client_secret, caches it until shortly before expires_in, and refreshes automatically (and once on a 401).

| Variable | Required | Notes | | ------------------------ | -------- | ---------------------------------------------------------------- | | ZOOMINFO_CLIENT_ID | yes* | DevPortal app client id (Okta 0oa…). | | ZOOMINFO_CLIENT_SECRET | yes* | DevPortal app client secret. | | ZOOMINFO_ACCESS_TOKEN | no | Pre-minted bearer token — testing only (ZI tokens last ~16 min). | | ZOOMINFO_BASE_URL | no | Override API base (default https://api.zoominfo.com/gtm). |

* Provide ZOOMINFO_CLIENT_ID + ZOOMINFO_CLIENT_SECRET or ZOOMINFO_ACCESS_TOKEN.

Token endpoint: POST https://api.zoominfo.com/gtm/oauth/v1/token (HTTP Basic, grant_type=client_credentials).

Tools

Connection / usage

| Tool | Description | | -------------------------- | ----------------------------------------------------------------- | | zoominfo_test_connection | Mint a token and hit the usage endpoint to verify credentials. | | zoominfo_get_usage | Current usage/limits (request/record buckets). Free — pre-flight. |

Data — Search (credit-free discovery → returns IDs)

| Tool | Endpoint | | --------------------------- | --------------------------------------------------------------------------------------- | | zoominfo_search_companies | POST /data/v1/companies/search | | zoominfo_search_contacts | POST /data/v1/contacts/search | | zoominfo_search_intent | POST /data/v1/intent/search (topics required — see zoominfo_lookup intent-topics) | | zoominfo_search_news | POST /data/v1/news/search | | zoominfo_search_scoops | POST /data/v1/scoops/search |

Data — Enrich (⚠️ credit-consuming: 1 credit per newly enriched record)

| Tool | Endpoint | | ----------------------------------- | ------------------------------------------------------------------------------------------------------- | | zoominfo_enrich_companies | POST /data/v1/companies/enrich | | zoominfo_enrich_contacts | POST /data/v1/contacts/enrich | | zoominfo_enrich_company_hierarchy | POST /data/v1/companies/corporate-hierarchy/enrich (pass hierarchy outputFields, e.g. familyTree) |

Data — Lookup (taxonomies, free)

| Tool | Endpoint | | ----------------- | ------------------------------------------------------------------------------------ | | zoominfo_lookup | GET /data/v1/lookup/{fieldName} — e.g. industries, job-titles, intent-topics |

Copilot / Context (read)

| Tool | Endpoint | | ---------------------------------- | ------------------------------------------------------- | | zoominfo_account_summary | GET /copilot/v1/companies/{companyId}/account-summary | | zoominfo_account_summary_ask | POST …/account-summary/actions/ask | | zoominfo_company_insights | POST /copilot/v1/companies/insights | | zoominfo_company_lookalikes | GET /copilot/v1/companies/lookalikes | | zoominfo_contact_recommendations | GET /copilot/v1/contacts/recommendations |

Account Summary and Insights require the corresponding Copilot entitlement on the ZoomInfo account; a 400/404 there typically means the account/token isn't provisioned for that surface, not a bad request.

Escape hatch

| Tool | Description | | ---------------------- | -------------------------------------------------------------------------------------------------- | | zoominfo_api_request | Call any read endpoint the typed tools don't cover. GET/POST only. Path is relative to /gtm. |

Request conventions (handled for you)

• POST bodies use the JSON:API envelope { "data": { "type": "…", "attributes": {…} } } with Content-Type: application/vnd.api+json.
• Pagination/sort are query params: pass page_number, page_size, sort as flat tool args — they're translated to page[number] / page[size] / sort (fully percent-encoded; the ZI gateway 400s on raw brackets).
• Search → collect ids → Enrich by id is the canonical, credit-efficient workflow.

Examples

// Find a company (free) — returns companyId
{ "tool": "zoominfo_search_companies", "arguments": { "companyName": "ZoomInfo", "page_size": 1 } }

// Enrich by id (1 credit unless already under management)
{ "tool": "zoominfo_enrich_companies",
  "arguments": { "matchCompanyInput": [{ "companyId": "344589814" }],
                 "outputFields": ["id", "name", "website", "revenue"] } }

// Intent — topics must be valid (see lookup intent-topics)
{ "tool": "zoominfo_search_intent", "arguments": { "topics": ["Adaptive Learning"], "page_size": 5 } }

// Buyer-intent topics taxonomy
{ "tool": "zoominfo_lookup", "arguments": { "fieldName": "intent-topics" } }

Build

cd servers/zoominfo && npx tsc
# → dist/servers/zoominfo/src/index.js

License

MIT