In our 2026 runs, Apollo's mixed_people/api_search cost 0 credits and people/bulk_match billed 1 credit per matched person even with both reveal flags off. Three parameters returned HTTP 200 and were ignored, and no endpoint returns a credit balance, so log credits yourself per job.
This is the reference we wished we'd had when we started calling Apollo's API in February 2026. It answers two developer questions: which calls spend credits, and how do you know a call did what you asked when it returns HTTP 200 either way? We ran the API until September 2026 on a plan with 11,000 credits a month, and everything below is what we saw on our key. Apollo can change billing, so re-run the checks at the end before you build on them.
Which calls bill
| Call | Credits in our runs | What came back |
|---|---|---|
POST /api/v1/mixed_people/api_search |
0 | First name, exact title, masked surname, has_email, Apollo person ID |
POST /api/v1/mixed_companies/search |
0 | Company matches by name |
POST /api/v1/organizations/bulk_enrich |
not logged by us | Firmographics and keywords, up to 10 domains per call |
POST /api/v1/people/bulk_match |
1 per matched person, even with reveal flags off | Full profile, current employer |
POST /api/v1/mixed_people/search |
none | HTTP 422, deprecated for API callers |
| Credit balance | none | No endpoint returns it |
People search is the free call to build around
mixed_people/api_search returned these fields at 0 credits:
| Field | What we got |
|---|---|
first_name |
Full |
last_name_obfuscated |
Fixed-width mask, e.g. Co***y |
title |
Full and exact |
organization.name |
Full |
id |
Apollo person ID, for a later paid call |
has_email, has_direct_phone |
Flags |
has_city, has_state, has_country |
Booleans only, no values |
email, last_name, linkedin_url |
Absent |
curl -X POST https://api.apollo.io/api/v1/mixed_people/api_search \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_KEY" \
-d '{
"q_organization_domains_list": ["example.com"],
"person_titles": ["revenue operations", "sales operations"],
"page": 1,
"per_page": 25
}'
person_titles matches substrings, so "operations" also catches "VP Operations" and "Operations Manager". The location fields are booleans, which means this call can't narrow people to one country on its own; filter on the company side instead. What the free search is good for, and how far it got us on a real account list, is in the Apollo review.
bulk_match bills with the reveal flags off
people/bulk_match turns person IDs or LinkedIn URLs into full profiles. We set both reveal flags to false because we only wanted titles and current employers, and it still billed 1 credit per person.
curl -X POST https://api.apollo.io/api/v1/people/bulk_match \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_KEY" \
-d '{
"reveal_personal_emails": false,
"reveal_phone_number": false,
"details": [{"linkedin_url": "https://www.linkedin.com/in/example"}]
}'
Lookups that matched nothing weren't billed. Our largest run, nearly 900 LinkedIn URL lookups, billed about 780 credits, and the gap is the misses. Budget for one credit per expected match, which is fair, but don't assume the flags make the call free.
Three parameters that return 200 and do nothing
| Parameter | What happened | Use instead |
|---|---|---|
organization_domains |
Ignored; the search ran across the whole database | q_organization_domains_list |
person_departments |
Ignored; we got the whole company | person_titles with role keywords |
similar_to_organization_domains |
Ignored; three unrelated seed domains returned identical totals | Nothing. There is no lookalike search |
The check for each one is the same. Run the query with and without the parameter and compare total_entries. If the totals match, Apollo dropped it.
curl -X POST https://api.apollo.io/api/v1/mixed_people/api_search \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_KEY" \
-d '{"organization_domains": ["example.com"], "per_page": 1}'
# compare total_entries with the same call using q_organization_domains_list
bulk_enrich keeps positions with nulls
curl -X POST "https://api.apollo.io/api/v1/organizations/bulk_enrich?domains[]=example.com&domains[]=example.org" \
-H "x-api-key: YOUR_KEY"
A domain Apollo can't match comes back as a null entry in the array, so positions stay aligned with your input. Guard for nulls before reading fields. The keywords field is much richer here than in mixed_companies/search, so don't compare a set enriched one way with a set enriched the other.
Tracking credits without a balance endpoint
Apollo documents a usage call (POST /api/v1/usage_stats/api_usage_stats), but it reports consumption against per-minute, hourly and daily rate limits, not credits. GET /api/v1/auth/health confirms the key works and nothing more. We found no call that returns a credit balance, so we kept a ledger:
- Read the credit count in the Apollo app before and after each job.
- Log the job date, endpoint, input count, matches returned and credits used.
- Count
bulk_matchmatches separately from reveals, since both bill. - Set a ceiling per job in code and stop the loop when the running count hits it.
Our plan ran dry in September 2026 and the ledger was the only warning we had. A recipe that gets usable addresses from the free search without spending Apollo credits is written up on what is waterfall enrichment.
Checks to run on your own key
- Call
auth/health, then run oneapi_searchwithq_organization_domains_listand notetotal_entries. - Swap in
organization_domainsand see whether the total jumps to the whole database. - Run 10
bulk_matchlookups with reveal flags off, two of them made up, and compare the credit count in the app before and after. - Send
bulk_enricha batch with one fake domain and confirm thenullentry. - Compare
person_departmentsresults with and without the filter.
Plan prices and credit allowances are on Apollo pricing.
Questions buyers ask
Is the Apollo API free?
People search is. mixed_people/api_search cost 0 credits in our 2026 runs and returned first names, exact titles, masked surnames and has_email. Company search by name also cost 0, while reveals and people/bulk_match billed credits.
How many credits does Apollo bulk_match use?
One per matched person, even with reveal_personal_emails and reveal_phone_number set to false. Lookups that matched nothing weren't billed.
How do I check my Apollo credit balance through the API?
We found no way to. The documented usage-stats call reports rate-limit consumption, not credits, so we read the count in the app before and after each job and kept our own ledger.
Why does my Apollo people search return the whole database?
The usual cause is organization_domains, which Apollo ignores while returning HTTP 200. Use q_organization_domains_list; person_departments and similar_to_organization_domains are ignored the same way.
Is mixed_people/search still available?
Not for API callers. It returned HTTP 422 as deprecated, and mixed_people/api_search is the working path.