13 tools, two access tiers.
Read tools — search, count, preview — are free and need no special permission. Account actions — reveal and export — consume credits and require an explicit OAuth scope plus workspace permission, so an agent can never quietly drain your data or budget.
count, preview, reveal, create_export, and create_list all accept the same filter set. Combine any of them; every filter narrows the same match.
Discovery (ChatGPT-compatible)
The two-tool contract ChatGPT and deep-research clients expect. Returns result handles and safe summaries only.
Search Argorant's safe coverage for a company, market, role, or location. Returns result handles — one for aggregate coverage, one for masked previews. Pass a handle to fetch.
Open a result handle returned by search. Returns the aggregate count, usage status, and up to 5 masked preview rows as a safe text summary. Never returns raw emails, phones, or profile URLs.
Read & preview
Free, read-only tools for sizing a segment and inspecting it without spending credits.
Check the connected workspace, role, access level, daily limit, and remaining usage. Call this first to see what the signed-in account is allowed to do.
Count matching contacts by any combination of the shared filters. Aggregate count only — the fastest way to size a market before committing credits.
Preview matches with masked initials, role, company, location, and department — no personal contact details. Defaults to has_email=true.
Reveal & export
Account actions that consume credits and require an explicit OAuth scope plus workspace permission.
Reveal real contact rows — emails, phones, profile URLs — for a small matching set (up to 100). Each email is verified live at reveal; only deliverable contacts are returned and billed. For large lists, use an export instead.
Create an async export from filters. Emails are verified at export time. Lists over one file become a tracked batch. Returns a job_id or batch_id to poll.
Download a completed export's CSV as text for small agent hand-offs. Large exports should be downloaded from the Argorant app.
Saved lists
Persist a segment once, then preview, export, or re-run it later.
Save a reusable list from filters or specific record IDs. Saving does not reveal contacts — exporting or revealing from the list consumes credits.
Fetch metadata and current size for a saved list.
Create an async export from a saved list, optionally skipping rows already exported. Returns a job_id to poll.
Job status
Poll long-running exports created by the tools above.
Get status and progress for a single export job.
Get status and progress for a multi-file export batch.
A complete agent flow
Size for free, inspect for free, then spend credits only on the rows you actually want.
// 1) Size the market — free, no scope
argorant_count_people({ industry: "saas", title: "vp sales", country: "DE" })
// → { "count": 8412 }
// 2) Inspect it safely — free, masked
argorant_preview_people({ industry: "saas", title: "vp sales", country: "DE", limit: 5 })
// → { "returned": 5, "results": [ { "initials": "J.K.", "title": "VP Sales", ... } ] }
// 3) Export verified rows — consumes credits, needs argorant:create_exports
argorant_create_export({ industry: "saas", title: "vp sales", country: "DE", limit: 5000 })
// → { "job_id": 48213, "status": "queued" }
// 4) Poll, then hand off
argorant_export_status({ job_id: 48213 })
// → { "status": "completed", "rows": 5000, "download": "…" }Every account action is tied to your signed-in Argorant workspace, its OAuth scopes, permissions, and credit balance — so the agent operates inside the same guardrails your team does.