apollo.

create an api key in the developer/api settings of your apollo account.

a free-tier platform key (daily quota) if you do not set your own — it

queries Apollo's SHARED database, the same for everyone.

says it resets at midnight — the response of a successful call also carries platform_quota

(used/limit/remaining) so you can stop before the refusal in the middle of a batch. Set your

own key to lift the limit right away, or for THIS lead: hunter_email_finder

(email) and kaspr_enrich_linkedin / fullenrich_enrich_linkedin (phone,

LinkedIn history) — different source, no credit burned on a call that

would fail anyway.

Apollo bills a reveal ON TOP of the match, whereas our platform counter can only

debit a bare match. This is true of the phone (~9 credits where a bare match costs

1) as well as personal emails (separate pot, plan-dependent scale — gap not measured,

so not debitable either). apollo_match_person keeps working without your key, it

simply returns no mobile, no direct dial, and no personal email.

on these tools, you need your own key. Not only to write: even listing or

reading them returns YOUR data (your contact book, your connected mailboxes, the content

of your sent emails, your call transcripts) — a pooled platform key

would expose that to any other oto user.

Integrations → API): a standard key authenticates but returns 403 on these three.

search and enrich companies and people, and spot hiring signals.

building uses: a search returns hundreds of obfuscated names, and this is how

they are revealed (300 people = 30 calls, not 300). ⚠️ **the batch saves

no credits** — apollo bills per PERSON, exactly like 10 single calls;

what it saves are calls and the rate limit. reveals (personal

emails, phones) there require your own key, as in single mode.

⚠️ the nested company record is lightened by default on apollo_match_person and

apollo_bulk_match: tech stack, funding rounds, subsidiaries and keywords weighed

91% of the payload — a single match came out at 60,000 characters and exceeded the output

limit of MCP clients. the name, domain, phone, headcount and industry

stay; full=True returns the raw, at the same price. in a batch, each record also loses

employment_history and account (the company record of your apollo CRM): without that, a

batch of 10 came out at ~86,000 characters.

⚠️ the only apollo action that does not return its result. apollo never returns a

mobile in the response: it verifies it on its side and POSTs it to a url, a few

minutes later. the immediate response only carries a request_id.

key, ~9 credits. webhook_url is mandatory on apollo's side: it is an

HTTPS url that YOU control (an n8n or make endpoint, your service) — **oto is not a

webhook receiver**, and does not see what lands there.

credit, for 30 days**. this is how the number comes back to the agent: you do not have

to read yourself what apollo posted. each record of people[] comes out lightened like the

reveal (employer tech stack, employment history, apollo CRM company

record removed — the numbers stay); full=True returns the whole envelope.

"Master" key), according to the apollo doc — same family of prerequisites as the contact

tools. to be verified on a real key: if it does not have it, the reveal still goes out

and the numbers arrive on your webhook, but apollo_reveal_phone_result will refuse.

credits are spent and there is nothing left to collect. after 30 days, the result

disappears for good.

treated as unauthenticated, and made idempotent.

do come back in the response (synchronous), on your own apollo key (same cost

rule as the phone). apollo withholds them for people in GDPR regions: an empty

result is an answer, not an outage.

SAVED on your side: title, email, phones, stage, lists, custom fields

database). ⚠️ The other source, often already paid for: apollo_match_person carries the

contact id NESTED at person.contact.id, as soon as the person is a contact on your

side. An id from apollo_search_people is a PERSON id and will be refused here.

The same contact_id is then used for apollo_sequence_contacts(op=add)

Apollo database, apollo_contact only sees what your team has already

saved. A person found but never saved has no contact id

apollo_match_person costs one and returns the shared record, not your values

is keyed by field id, never by name. For a picklist, the

value to write is the option id, not its label

interface (useful when the account belongs to the client). ⚠️ For a long text —

a hook, a paragraph — use field_type="textarea": string is

capped at 120 characters and Apollo truncates without saying anything. A field already

carrying this name makes the creation be REFUSED rather than creating a namesake

Integrations → API); a standard key authenticates but returns 403

mailboxes / send schedules), to call before creating a sequence or enrolling

contacts in it

sequence

contacts, consult their activity. add starts an automated campaign to real

people — dry_run available

nothing goes out. search/content return YOUR sent emails (body included), not a

shared database

the one your client asks you to confirm

recordings. The credit cost of a get depends on the presence of AI insights,

unpredictable before the call

this connector accepts multiple comptes: each stored credential becomes a named account (one name per compte), at your level, your team's or your org's.