Apollo review (2026): easy to start, limited by one dataset

What is Apollo?

Apollo is a B2B contact database with a sales engagement suite built around it. You search its 240 million contacts by title, company size, industry, or technology, reveal emails and phone numbers, and send them into sequences, calls, and meetings without leaving the app. Tim Zheng and Ray Li founded it in 2015; Matt Curl took over as CEO in February 2026.

The appeal is consolidation. A founder or a three-person sales team gets a database, a sequencer, a dialer, a meeting scheduler, and a Chrome extension on one bill, starting at nothing.

Where Apollo's data comes from

Apollo builds its database from four sources, according to its own data overview: a contributor network of more than two million users who share email and calendar signals, engagement data from its own tools (replies and bounces), public web crawling, and third-party providers. Paid plans add a waterfall that checks outside partners for missing data.

That mix is the weakness. A database built from its own users is good where those users sell, which in our experience means US tech companies, and thin or stale everywhere else: European contacts, small businesses, uncommon job titles, and anyone who changed jobs since the last signal came in. When Apollo is wrong or empty, there is nothing behind it. No second provider gets asked.

Apollo's own numbers don't agree with each other either. Its data page claims 98% email accuracy. Its help center article on email verification, updated in September 2026, describes a seven-step process and "a 91% accuracy rate". At 91%, roughly one email in eleven is wrong before you send. We haven't benchmarked Apollo ourselves, so test it on twenty accounts you know well; that is usually enough to see the gaps.

What Apollo costs in practice

Apollo prices per seat, and every seat comes with credits. An email costs 1 credit and a mobile number 8, charged only when Apollo finds something. A Basic seat at $49 a month billed annually includes 2,500 credits.

The sticker price is low; the shape of the bill is the problem. Every person who logs in is a seat, so a five-person team on Basic pays $245 a month before buying a single extra credit. Credits don't roll over, purchased ones included, so a seat that uses half its credits pays double for each contact. A team that needs more data but not more people either buys add-on credits (1.5 to 2.5 cents each) or pays for seats nobody uses. The Organization plan needs three seats on an annual contract.

A fixed data model

Apollo's records are Apollo's fields: 65+ data points per person or company, defined by Apollo. Lists filter that database; they are not tables you can extend. There is no column that asks a second provider for a missing phone number, runs an AI prompt over a company's website, or scores an account with your own logic. Anything beyond a contact lookup happens in another tool, after an export.

The API and agents

Apollo has an MCP server at mcp.apollo.io/mcp and a REST API, both on every plan. An agent in Claude, ChatGPT, or Cursor can search the database, enrich contacts, add them to sequences, and update the CRM.

The limits show up as soon as you use them for real work. General API endpoints allow 400 requests an hour and 2,000 a day on Basic and Professional, and 600 a day on Free; Organization raises the daily cap to 6,000. Enrichment and search endpoints get separate, higher limits on paid plans, but an agent that looks up accounts, people, and sequences in one run touches several endpoints and hits the general ceiling first. And whatever the agent does, it can only read Apollo's own fields from Apollo's own dataset.

Apollo vs pipe0

pipe0 does not replace Apollo's sequences or dialer; it hands enriched lists to outreach tools. Where they overlap is data, and there the difference is structural. Apollo answers every lookup from one database. pipe0 runs waterfalls across 50+ providers and stops at the first that answers, so a miss in one dataset becomes a second and third chance instead of an empty cell. You pay only for the provider that found the result, and there are no seats: plans start at $49 a month for the whole team, and we measure about 14 cents per mobile number found.

The data model is open, too. A pipe0 sheet holds up to 2 million rows, and any column can run an enrichment, an AI prompt, or your own logic, on a schedule, into HubSpot, Salesforce, or Attio. The same blocks are an API and an MCP server built for agents running at volume.

Many teams use both: pipe0 to find and verify the data, Apollo to send. The full comparison covers the differences.

Who should buy Apollo

Buy Apollo if you are a small team selling into US tech, you want one tool for finding people and contacting them, and speed of setup matters more to you than coverage outside that market. Start on the free plan and check its coverage of twenty accounts you know well before you pay.

Look elsewhere if data quality outside US tech decides your pipeline, if you enrich over the API or with agents at any volume, or if your enrichment goes beyond a contact lookup. A waterfall, or a table like Clay, will return more.