Configuration
Create an API key with endpoint
permissions for the resources you select. Requests use
https://api.apollo.io/api/v1
and send the key in the x-api-key header. Connection testing searches contacts,
so contact-search access is required even when selecting other resources.
The defaults use credit-free endpoints and do not require a Master API key.
Optional resources need their own endpoint permissions and any associated plan
features.
calls and email_stats require a Master API key. Consult the
Apollo API reference for the selected
endpoints. conversation_details consumes one credit for each conversation with
AI insights; conversations without AI insights cost no credits.
Resources
Contacts, accounts, and sequences are selected by default. All other resources are optional. Selecting a child also reads its parent to discover IDs, but only selected resources are written.
Each record resource emits one row per Apollo ID. Nested account data, custom
field values, lists, recipients, and other structures remain JSON. Fields not
projected into columns are retained in
raw.
email_content fetches stored HTML for sent messages; Apollo omits IDs that do
not match a sent email. email_stats emits one row per email with its activities
in a JSON array. contact_sequence_activity emits one row per contact containing
at most the latest 50 events across its sequences. Apollo provides no pagination
for older events. conversation_details includes transcript and recording
metadata; it does not download recordings. Usage resources each emit one snapshot
row, retaining the endpoint-keyed API usage object in raw.
This connector reads your saved records rather than searching Apollo’s global
people and organization database. Global prospect searches, enrichment, company
news and job postings, custom analytics report queries, asynchronous exports,
webhook-result polling, and write operations are not included. Separate CRM
record detail lookups and send-status polling are not exposed; list responses
supply those records.
Modes
All resources support full reads. The documented CRM searches lack an updated-since filter. Sorting byupdated_at does not restrict a read to changed
records. Notes filter by creation date, and call, email, and conversation date
filters do not reliably capture later edits, so these resources also use full
reads.
credit_usage and api_usage have no stable row key. Use full-replace writes
for their snapshots. Full reads do not emit deletion events; use full-replace
writes when the destination must remove records no longer returned by Apollo.
Behavior
- Auth: Static API key authentication. This connector does not implement Apollo’s partner OAuth authorization or token refresh flow. Permission and plan errors fail the read.
- Pagination: Contacts and accounts send page numbers and a page size of 100
in JSON bodies. Conversations use body pagination with 25 results per page.
Other paginated searches use query page numbers with 100 results per page.
Notes use
skipandlimit. Reference lists and snapshots use the endpoint’s unpaginated response. - Search limits: Apollo caps contact, account, and outreach-email searches at 50,000 records. The connector does not partition searches to bypass that cap. Larger datasets cannot be treated as complete exports. See contact search, account search, and outreach email search.
- Parent reads: Optional email content, email statistics, contact activity, and conversation details make one request per parent record. Their coverage is limited to records discoverable through the parent search. Repeated full reads of conversation details can consume credits again.
- Rate limits: Apollo applies per-endpoint minute, hour, and day limits that
vary by plan. The shared HTTP retry policy handles retryable failures, but the
connector does not coordinate plan-specific quota budgets.
api_usageexposes current usage and limits when your key has access. - Consistency: Page-based searches are not point-in-time snapshots. Records changing during a read can affect page membership and ordering.