
The batch ad-audience endpoint gets a lot done in one call. We sent thousands of sales-leader contacts from Salesforce (LinkedIn URLs, with work email as a fallback) in a single batch. Within an hour, 92.6% came back matched, with SHA-256 hashed identifiers: several emails per person on average, plus phones, name and location. Because those hashes don’t depend on the ad network, one run gave us upload files for Meta, Google, TikTok, X and Reddit. We didn’t have to pay once per platform.
Pricing is fair. You pay for profiles found, not rows submitted: misses are refunded and failed requests are never charged. The work-email-from-LinkedIn endpoint follows the same rule. We ran it on contacts with no email in our CRM and found 89%, nearly all on the person’s current company domain, and paid nothing for the misses. It also checks where the person works today, which kept us from attaching an email from someone’s old job to the account.
The API is easy to build on. One API key header, a free balance check, and the credit cost and remaining balance on every response made it simple to put spend guardrails into our automation. Their MCP server took a single command to add to Claude Code, which put about two dozen enrichment and research tools straight into our AI workflow.
The team was quick to help. They added their CTO to our shared Slack channel on day one. Review collected by and hosted on G2.com.
Small things, mostly things to know before you build:
The default rate limit is 1 request per second per account, so paging large batch results needs some backoff.
The full OpenAPI spec is behind the web login rather than the API key.
The ad-audience output is hashed only, which is right for ads, but readable emails come from separate endpoints at separate cost. The personal-email lookup returns one best address, not every one on file.
None of this slowed us down for long, and the docs and their team covered the rest. Review collected by and hosted on G2.com.