The alternative to a B2B contact database is not a bigger database
Every contact database is a snapshot. It is bought, imported, exported into a sequencer and then starts dying the moment it lands — because the people in it move, and their mailboxes get retired. TAPAC takes the opposite approach: it does not store a copy of the market, it queries the live web (company websites, Telegram channels, Discord servers) when you ask, and verifies each mailbox over SMTP in that same call.
What "the database decayed" means in numbers
- 23% of business contacts change jobs every year (ZoomInfo, 2025).
- 40% of email addresses are dead within two years (NeverBounce).
- Fresh, scraped data outperforms stale database records by 42% (Harvard Business Review, 2024).
- Community-sourced leads convert 30% better than cold outreach to a purchased list (McKinsey, 2025).
- TAPAC's own returned-contact bounce rate is 2–5%, against the 10–35% typical of static databases.
None of that is an argument against databases existing. It is an argument against paying a subscription for the part of a database that has already expired.
Two models, side by side
| TAPAC (sourced per request) | Stored contact database | |
|---|---|---|
| Where the data comes from | Live sourcing: company sites, Telegram, Discord | A copy crawled and stored earlier, resold |
| What verification means | SMTP handshake at request time, per address, returned with the contact | Platform-level check, usually at import time |
| Age of the record you send to | Minutes | Whatever the last refresh left behind |
| Typical bounce | 2–5% | 10–35% |
| Pricing unit | $0.10–0.50 per verified contact, pay-per-use | Seat licences plus credits — e.g. Apollo Basic $49/user/month billed annually, Lusha Starter $49.90/month, Hunter from $34/month |
| Cost for a 3-person team | Only the contacts actually used | Three seats from day one, whether or not they source |
| Access for agents | MCP server (npx -y @tapacapi/mcp) + REST API + OpenAPI 3.1 | REST API, browser extension, no MCP server |
| Free entry | 100 free searches, no card | Credit-limited free tiers (Hunter: 50 credits; Lusha: 40 credits/month as published in plan comparisons) |
Competitor figures are published list prices as of writing (apollo.io/pricing, lusha.com, hunter.io) — re-check before buying. TAPAC figures come from the product's own documentation.
When a database is still the right buy
If your job is to search a very large universe by dozens of firmographic filters and you are happy to pay per seat for that convenience, a database does it. TAPAC is narrower by design: give it a target (industry, roles, size, geography, source) and it returns verified people you can write to today.
Run it in a minute
npx -y @tapacapi/mcp
That registers the TAPAC MCP server for Claude, Cursor, Codex, Windsurf or any MCP client and exposes two tools — tapac_find_contacts and tapac_status. Prefer HTTP: POST /v1/contacts/search with a Bearer key from /get-key. Over HTTP with the JS SDK:
npm install @tapacapi/sdk
import { TapacClient } from "@tapacapi/sdk";
const tapac = new TapacClient({ apiKey: process.env.TAPAC_API_KEY });
const {{ contacts }} = await tapac.findContacts({{ industry: "SaaS", job_titles: ["VP Sales"], limit: 20 }});
Where to go next
- REST API quickstart — endpoints, Bearer auth, sandbox header, OpenAPI 3.1 spec at
/openapi.json. - Get a free API key — 100 searches, no credit card.
- Verify email before sending · Lead generation MCP server · Install the MCP server
- Compare TAPAC with Apollo, ZoomInfo, Hunter and Lusha · Apollo alternative · ZoomInfo alternative · Hunter alternative · Lusha alternative