Connect Claude, GPT or your own agent — search 565.8M contacts over MCP
Argorant
Argorant
Get Started
Agent-native · MCP · API · CLI

Your prospecting, fully agent-run.

Connect Claude, ChatGPT or your own agent to 565.8M contacts. It counts your market for free, previews the people, and buys exactly the leads you approve - pay per lead, no plan required.

See agent pricing
Watch the 90-second setup
From copied prompt to your first verified list.
ClaudeClaude CodeChatGPTCodexGrok BotGrok BuildGrokCursorPerplexityOpenClaw
Pay per lead

No subscription. Your agent buys leads one at a time.

Top up a prepaid wallet, and your agent spends it credit by credit: $0.035 per verified contact, counts and previews always free. Auto top-up keeps long runs unattended. Doing volume? Subscriptions bring the price down to $0.022 - your agent spends plan credits through the exact same tools.

  • Wallet top-ups from 500 credits
  • Every spend logged and approvable
  • Same account as the UI - credits, lists, exports
Full agent pricing
Wallet
500 credits$17.50
1,000 credits$35.00
3,000 credits$105.00
$0.035 per credit · one verified contact = one credit
capabilities

Everything your agent can do with your market

Six governed capabilities behind one connector. Free where it should be free, metered only where value leaves the platform.

Count any market

Real population sizes for any industry, role, size band and geography - free, unlimited.

33,056
CFOs · Europe · live count
Search & preview

Redacted previews of the actual matching people, so the agent can sanity-check before spending.

A** P***** · CFOpreview
M** K**** · CFOpreview
S** L******* · CFOpreview
Reveal on approval

One credit per verified contact, only when you or your agent says go. Every spend logged.

Reveal the top 5?
Done - 5 contacts, 5 credits. 495 left.
Verified exports

CSV exports run a live mailbox check at download - lists arrive deliverable.

j.weber@meyer-gmbh.devalid
info@alt-domain.ioinvalid
office@bau-ag.decatch-all
Saved lists & re-runs

ICPs live as saved queries the agent re-runs whenever positioning changes.

DACH CFOs 200+4,711
UK accountants45,627
US contractors107,249
Spending guardrails

Counts free, spends explicit, scopes per key - an agent can be curious without being expensive.

Export all 4,711?
That costs 4,711 credits - confirm first.
Two-minute setup

Connect the client you already use.

Remote MCP server URL
https://mcp.argorant.com/mcp
1
In Claude, open Settings → Connectors → Add custom connector.
2
Paste https://mcp.argorant.com/mcp as the remote MCP server URL.
3
Complete the Argorant OAuth sign-in and enable the connector.
Full Claude guide
What it looks like

One question in. A verified list out.

“Find 200 verified CFOs at German manufacturers.”
→ count_matches · 8,412 found
→ create_export_link · 200 requested
cfos_de.csv · preview
Katrin VossCFO · Hartmann Maschinenbauvalid
Jonas BrendelCFO · Weiland Kunststofftechnikvalid
Sabine KrügerCFO · NordStahl Gruppevalid
Markus EhlertCFO · Feinwerk Antriebevalid
198 valid · 12,40 € · export ready
Counts and previews are free — credits are only spent on the export, and only for rows verified valid.
Thirteen governed tools

A tool catalog built for agents, not a raw query console.

Each tool does one job well — search, count, resolve a market, reveal a contact, or hand a finished export to the workspace owner.

Full tools reference
01

Connect

Anything that speaks MCP: Claude (Desktop, Cowork, and Code), ChatGPT connectors, Cursor, Codex, Perplexity, Manus, OpenClaw, Hermes, and any custom runtime. You decide which tools an agent may call.

MCP endpoint
https://mcp.argorant.com/mcp
03

Verify

Exports run a live SMTP check at download time, and uploaded lists can be verified through the same engine via the API. Verdicts come back per address: valid, invalid, catch-all, or unknown.

Live SMTP check · founder@nordwind-sys.com
Syntax & domainpass
MX records foundpass
SMTP handshakepass
Mailbox accepts RCPTpass
Catch-all probenot catch-all
Resultvalid · deliverable
04

Export

Exports are verified at the moment of download. Credits are only spent when contacts are revealed or an export is created — every spend is logged.

app.argorant.com/exports
DACH - Manufacturing CFOs
Leads · verified at export · skips prior exports
done
5,234 / 5,420 checked live96.6% deliverable
valid · 4,981catch-all · 253invalid removed, not billed
Download CSVPush to sequencer
API + MCP on every paid plan.

Public counts and redacted previews are available before checkout. API and MCP rate limits scale with your paid plan or agent wallet.

See pricing
Free10 req/min · search, count & preview tools only
Starter30 req/min · full API + MCP
Pro120 req/min · full API + MCP + webhooks
Scale600 req/min · priority queue + webhooks
Build on it

A REST API and signed webhooks for production.

Everything an agent does over MCP is also a clean REST endpoint. Signed, retried webhooks fire on exports, verifications, and low credits — no polling.

POST /v1/exports
curl -X POST https://api.argorant.com/v1/exports \
-H "Authorization: Bearer ak_live_4f2k…" \
-H "Content-Type: application/json" \
-d '{
"industry": "manufacturing",
"country": "DE",
"grades": ["valid"]
}'
Response · 201 Created
{
"id": "exp_9f2c",
"total": 8412,
"verified": 6930,
"export_url": "https://argorant.com/e/exp_9f2c"
}
Webhook eventsPOST → your endpoint
export.completed5,234 rows verified · leads.csv ready14:32:08
verification.batch_finished1,000 checked · 94% valid14:31:42
credits.lowworkspace at 12% remaining14:29:05
export.createdjob 48214 queued · 1,180 rows14:26:51
signed · retried with backoff · no polling
Workspace-scoped keys and OAuth grants — agents act under your account, never around it.
Tool allowlists: you decide what an agent may call.
Per-call audit log of every search, reveal, and export.
Masked previews by default; reveals and exports metered in credits.

Frequently asked
questions

Agents, answered

Free actions, credit metering, rate limits, and how access stays governed.

Still have questions? Talk to us

No. Pay-per-lead runs on a prepaid wallet: top up from 500 credits at $0.035 per credit and the agent spends it one verified contact at a time. Subscriptions are cheaper per credit ($0.022 on Starter) once you do real volume - both work through the same API and MCP.
Anything that speaks MCP: Claude (Desktop, Cowork, and Code), ChatGPT connectors, Cursor, Codex, Perplexity, Manus, OpenClaw, Hermes, and any custom runtime. There's also a REST API and a CLI on npm for everything else.
Counts, searches, and masked previews cost zero credits — on every plan, including Free. CLI counts don't even need a key. Credits are only spent when contacts are revealed or an export is created.
Search, count, and preview tools are free. Reveal and export tools draw from the same workspace credit pool as the app — one credit per deliverable contact (valid, or catch-all/risky if opted in), zero for invalid. Every spend is logged.
Free: 10 req/min on preview-class tools. Starter: 30 req/min. Pro: 120 req/min. Scale: 600 req/min with a priority queue. API and MCP are included on every paid plan — limits scale with the tier, never gated to a single plan.
Yes. Exports run a live SMTP check at download time, and uploaded lists can be verified through the same engine via the API. Verdicts come back per address: valid, invalid, catch-all, or unknown.
Yes — published on npm as 'argorant'. Run npx argorant count "..." with zero setup, or npm install -g argorant. Full reference at argorant.com/docs/cli.
Keys and OAuth grants are workspace-scoped. You control which tools an agent may call, plan limits cap throughput, previews are masked by default, and every action lands in an audit log your admin can review.
By agent

Setup guides for the agent you already use.

Claude Cowork

Anthropic's agentic workspace

Hermes

Autonomous agent runtimes

Codex

OpenAI's coding agent

Perplexity Computer

Perplexity's agentic research computer

ChatGPT

ChatGPT agent mode & connectors

Claude Code

Anthropic's coding agent

Cursor

The AI code editor

Manus

The general-purpose autonomous agent

Your agent could be building lists tonight.

“We run cold outbound for a dozen clients. Argorant replaced two tools and the bounce complaints just stopped.”Founder of a bootstrapped outbound agency · Lisbon, Portugal
Connect free
npm: argorantCLIOpenAPI / REST docsMCP docsTools reference
Loved by revenue teams
“We've gotten 16.81% in replies after working with you guys. Can only recommend you.”
Jannik Linke, Founder & CEO, Clementine Media · Germany