How to Choose an API-First or Platform-First B2B Data Strategy (for a Technical GTM Team)

TL;DR

• Go API-first if your data feeds code: enrichment on signup, scoring, your own tables, agents. Go platform-first if humans work lists in a UI every day.
• The test takes about 30 minutes: one free count, one 5-row search, one enrichment, one export price preview.
• It spends about 6 LeadOcean records, well inside the 1,000 on the Free plan.

Prerequisites

• A LeadOcean account on the Free plan: https://app.leadocean.io/register. No card.
• An API key from the app, exported as $LEADOCEAN_API_KEY.
• curl and jq.
• Optional: an MCP client such as Claude Code, to run the same count as a tool call.
• A target list in one sentence. Example: VPs of Engineering in the US.

Step 1: Size the list before you pick anything

Your list size and refresh rate decide the architecture, so measure them first. A count is free on every plan.

curl -s -X POST "https://api.leadocean.io/v1/people/search?count=true" \
  -H "x-api-key: $LEADOCEAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jobLevel":["VP"],"jobFunction":["Engineering"],"country":["US"],"limit":1}' | jq .meta

count=true goes in the query string, not the JSON body; limit 1 goes in the body. The body is flat, and every filter is an array. meta.total is capped at 100,000.

Write down two numbers: rows now, and rows you will pull each month. Under a few thousand rows a month that nobody edits in code, a platform is fine. Above that, or when code reads the result, lean API.

Step 2: Pull a 5-row sample through the API

An API-first flow is a loop your code owns: filter, page, store. Test the loop with five rows.

curl -s -X POST "https://api.leadocean.io/v1/people/search" \
  -H "x-api-key: $LEADOCEAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jobLevel": ["VP"],
    "jobFunction": ["Engineering"],
    "country": ["US"],
    "limit": 5
  }' | tee page1.json | jq '.meta'

This spends 5 records. Page size runs from 1 to 100. To get the next page, pass meta.nextCursor back as cursor.

One search pages to 10,000 rows. When meta.depthCapped is true, there is no nextCursor. Narrow the filters instead of walking further.

Check what your code would need to handle here: pagination, the depth cap, retries on 429. If that list feels like a chore, that is a signal for a platform. If it feels like twenty lines, stay API-first.

Step 3: Enrich one person by ID

Enrichment is the call most pipelines hit first. It takes exactly one of linkedin_url, email, phone or person_id.

PID=$(jq -r '.data[0].meta.source_ids.person_id' page1.json)

curl -s -X POST "https://api.leadocean.io/v1/people/enrich" \
  -H "x-api-key: $LEADOCEAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"person_id\":\"$PID\",\"reveal_email\":true}" | jq 'keys'

This spends 1 record. A person we hold nothing for returns 404 and costs nothing. Since October 2026 the response also carries career history (resume_data): past roles, education, certifications and awards.

There is no name plus domain lookup. Search by domain first, then enrich by person_id.

Step 4: Price the platform path with an export

Platform-first means a person picks filters in a UI and downloads a file. LeadOcean's app has this on its Exports page: filter builder, column picker, record price before you start, CSV download.

1. Sign in at https://app.leadocean.io and open Exports.
2. Set the same three filters: job level VP, function Engineering, country US.
3. Pick columns. Email columns are email_N_address and email_N_status.
4. Read the price preview. Do not download on the Free plan: bulk CSV export is not included there.

The same job over the API is POST /v1/exports. It takes up to 50,000 rows a request and 500,000 a day.

LeadOcean is not a rep workspace. It has no sequencer, dialer or CRM. If your team needs those, pair LeadOcean with a tool that has them. To work lists in a table instead, see LeadOcean + Clay.

Step 5: Apply the decision rule

Score your answers from steps 1 to 4. Do not average them. One hard "code reads this" outweighs three soft platform wins.

My rule: three or more rows in the left column, go API-first. Two or fewer, start with exports and add the API when a script appears.

Most technical teams end up hybrid. Exports for the one-off list, the API for anything a user triggers. Both draw from the same data on the same plan, so you do not choose once.

On cost, LeadOcean has one data set on every plan. Free is 1,000 records for life. Solo is $97 a month (50 requests a second). Agency is $497 a month (200 requests a second, 5 keys). Platform is $1,997 a month for embedding and reselling. Paid plans use a fair-use pace, not credits. See pricing.

What you get

Step 1 as an MCP count on 2026-10-21, using leadocean_count_leads with jobLevel VP, jobFunction Engineering, country US. The count defaults to mailable addresses (verified, catch_all_valid or catch_all):

{
  "total": 25784,
  "totalIsExact": true,
  "totalAll": 54103,
  "distinctCompanies": 11896,
  "credits": 0
}

So 25,784 mailable people at 11,896 companies, out of 54,103 matching people with no email condition. It cost 0 records.

Step 2 returns this shape (response shape from the API reference, placeholder values):

{
  "data": [
    {
      "profile_data": { "profile_full_name": "Jane Doe" },
      "contact_data": { "has_email": true, "email_status": "verified" },
      "resume_data": { "<current role and company>": "<values>" },
      "meta": { "source_ids": { "person_id": "<digits>" } }
    }
  ],
  "meta": { "count": 5, "limit": 5, "credits": 5, "nextCursor": "<cursor>" }
}

For more on what these APIs return, read what a B2B data API is.

Troubleshooting

FAQ