From October 1, WhatsApp charges for every business reply. Read the migration playbook →

Platform — Tools & automation

Agents that do things, not just say things

Give your agent real capabilities: call any HTTP API, drive a real browser for the sites that don't have one, search the web, and take payments — each tool metered and permission-scoped.

HTTP tool builder

Wrap any REST API as an agent tool with templated inputs — no code, SSRF-guarded by design.

Browser automation

A managed cloud browser completes multi-step website tasks — form fills, lookups, bookings — and reports back with results and PDFs.

Web search

Agents can search the live web when your knowledge doesn't cover the question, with sources attached.

Payments

Agents can request payment mid-conversation — checkout links generated and reconciled through the platform's billing.

Tool library

Tools are workspace-level and reusable across every agent and MCP server you publish.

What teams build with it

Three jobs this does the day you turn it on.

01

Book, quote and check — not just answer

Wrap the endpoints behind your booking or quoting flow so the agent finishes the job inside the conversation instead of handing over a link and hoping.

02

Work the portals that have no API

A managed cloud browser signs in, fills the form and reads the result back — for the supplier portals and legacy systems that were never going to ship an API.

03

Take the payment in the chat

The agent can raise a checkout link mid-conversation and confirm once it clears, so the sale does not wait on a follow-up email.

Where you set this up

The same build screen every feature is configured from — this row is the one you would touch.

build.supercognit.com
System prompt
You are the concierge for a 132-villa resort. Answer from the handbook, never invent a price…
Knowledge
handbook.pdfpricing.mdyourbrand.com
Tools
⚙ check_availability⚙ browser_task⚙ web_search
Channels
WidgetWhatsAppMCPTelegram
PublishNo code. No deploy step.

Under the hood

The details you will want before you commit.

HTTP toolsAny REST endpoint, templated inputs, SSRF-guarded by design
BrowserManaged headless sessions with an encrypted credential vault
Long tasksBrowser runs continue in the background and announce the result
MeteringPer-call metering with per-seat and workspace cost caps
ReuseA workspace tool library shared across agents and MCP servers

How it works

01

Build or pick

Create an HTTP tool from any API, or enable built-ins like browser tasks and web search.

02

Scope it

Each tool is permission-scoped and metered per call, with cost caps you control.

03

Let the agent work

The agent plans multi-step tool use on its own and reports what it did.

Frequently asked questions

Not for HTTP tools. Paste the endpoint, name the inputs and describe when it should be used; the builder handles the request, the auth header and the response.

Credentials live in an encrypted vault, are injected at run time and never appear in the transcript or in the agent's context.

Every call is metered, and both per-seat and workspace-wide cost caps are enforced — when a cap is reached, calls stop rather than quietly spending.

Yes. The agent plans multi-step tool use on its own, and the conversation shows which tools ran and what each one returned.

Ready to ship tools & automation?

Free for 7 days — publish your first product this week.

Start building