SellSea / Live product / 2026

SellSea

An outbound AI voice platform that turns a sales brief into real, reviewable conversations.

Visit the live product
Contribution
Strategy, identity, product design and engineering
Delivery
Marketing site, SaaS product, API and voice agent
Status
Live at SellSea.ai
SellSea live call demo displayed on a laptop in a dark audio studio
Brand, marketing, product and voice infrastructure designed as one connected experience.

A sales platform that makes its own first call.

Outbound work often breaks across prospecting tools, spreadsheets, diallers and handwritten follow-up. SellSea was designed as one operating loop: define how the agent should represent the business, give it the right contacts, run the calls and make every outcome useful to the human team.

Let people hear the product, not just read the promise.

Trust is the hardest part of an AI calling product. The public experience therefore centres on a real call-me demo. A visitor enters a verified UK or Spanish mobile number, receives a call and hears the qualification flow on their own phone.

That interaction turns an abstract technical claim into a direct product experience before asking for an account or card.

SellSea website showing the live AI caller and mobile number field
The live call-me demo on SellSea.ai.

One path from a prospect list to the next useful action.

The interface hides a multi-service system behind five clear stages. Configuration stays as data, so every organisation can deploy its own sales agent without rebuilding the voice worker.

  1. 01

    Find or import

    Build a prospect list in-product or bring an existing CSV into the workspace.

  2. 02

    Configure the agent

    Set the spoken name, company context, pitch, goal and qualifying questions.

  3. 03

    Run the campaign

    Choose the contacts, calling window and schedule, then start or pause the work.

  4. 04

    Hold the call

    The voice worker listens, responds and keeps the conversation moving in real time.

  5. 05

    Review the outcome

    Each call returns a transcript, summary, outcome and clear next action.

Make the work visible at every stage.

Campaign status, live notifications, scheduled work and call outcomes use the same compact visual language. The product should feel calm while a complex system is working underneath.

SellSea workflow interface showing campaign and notification states
Campaign work and real-time outcomes.
SellSea capability grid covering conversations, objections, meetings and call summaries
A direct explanation of the product system.

A signal that feels active without becoming noisy.

The mark is built from offset strokes that read as voice, movement and exchange. Electric blue creates an immediate product signal, while deep navy and pale cyan give the system enough range for operational screens and expressive marketing.

SignalVoiceLiveInkAir
SellSea blue wordmark
SellSea offset signal mark

More than a polished voice demo.

The work spans the customer-facing experience and the less visible systems needed to run it safely as a multi-tenant SaaS product.

Live call demo

Prospects can hear the agent on their own phone before creating an account.

Agent builder

Business context and scripts are stored as versioned workspace configuration.

Prospect Finder

Search, enrich and promote relevant companies into the sales workflow.

Campaign operations

Import contacts, dispatch calls and track progress from one working view.

Call intelligence

Transcripts, summaries and outcomes make every conversation reviewable.

Operational controls

Calling windows, plan limits and organisation-owned caller IDs constrain dispatch.

Usage and billing

Stripe subscriptions and a minute ledger connect product access to real usage.

Workspace model

Organisations, memberships and permissions keep product data tenant-specific.

Design decisions carried into the real system.

SellSea is built as a TypeScript monorepo. React powers the public site and authenticated product, Express and PostgreSQL manage product data, and a stateless LiveKit worker receives each call's configuration at dispatch.

That architecture keeps the voice engine reusable while the agent's identity, script, contact and business context remain specific to each workspace and call.

  • React 19 and Vite
  • Express and TypeScript
  • PostgreSQL and Prisma
  • LiveKit voice agents and SIP
  • Stripe billing and usage metering
  • Sentry and structured logging

A live product that can introduce itself, make the call and show the work afterwards.

Experience SellSea