0seat
The AI layer for Chatwoot and WHMCS

Your agents stop writing replies. They approve them.

Suspensions, invoices, DNS, migrations: the same questions come back all day. 0seat plugs into your Chatwoot inbox and WHMCS, learns from your resolved conversations and prepares the reply to every new message, using the customer’s real account. Your team reviews, sends, and puts the time saved into sales, migrations and the customers who matter.

  • Voice notes and screenshots understood
  • Automatic replies, category by category
  • Real-time voiceComing soon
Works with
  • Chatwoot, self-hosted or cloud
  • WHMCS, read-only
  • WhatsApp, via the official API

The same work, a hundred times a day

At a hosting company, most messages need no expertise. They need the right information at the right time.

The same questions

Invoices, suspensions, DNS, email, migrations: your agents rewrite the same reply, customer after customer.

Back and forth in WHMCS

For every message: open the account, find the service, check the due date, and only then reply.

Customers kept waiting

At peak hours the queue grows, and the requests that matter wait behind the simplest ones.

AI on your free Chatwoot, with the WHMCS account on top

Chatwoot’s own AI is not part of its free edition, the one most hosting companies run themselves. 0seat brings AI to it, with no plan change and no per-agent fee.

  • Chatwoot Community alone

    USD 0

    No AI: your agents write every reply themselves.

  • Chatwoot with Captain

    USD 19 to 99 per agent per month

    Chatwoot’s AI, reserved for paid plans, with limited credits on the cloud. No WHMCS integration advertised.

  • Chatwoot Community with 0seat

    $19 excluding tax per workspace per month

    As many agents as you like, the model at measured, capped cost, and the customer’s WHMCS account in every reply.

Chatwoot prices read on October 4, 2026 on its pricing page, cloud and self-hosted editions.

Why the free edition has no AI, and what 0seat adds

7 to 8 out of 10 customer requests get a correct reply on the first try

Internal measurement on 10 real requests at a hosting company, October 2026. A measurement on a larger volume is planned.

What 0seat does for your team

From the knowledge base to the handoff, every step is visible and adjustable in the console.

Knowledge base

It learns from your resolved conversations

0seat reads your history, extracts the questions and answers that worked, and proposes them to your team. No entry is used before it has been approved.

  • Extracted from your conversation history
  • Approved, edited or rejected by your agents
  • New entries proposed every night
Explore the knowledge base
0seat console, Knowledge base tab: 12 entries to review, 148 approved, one entry “My domain name stopped responding this morning” with its answer and the Approve, Save without deciding and Reject buttons.
Prepared replies

The reply is ready before the agent opens the conversation

For every new message, 0seat writes the reply as a private note, visible to your team only. The agent reviews it, edits it if needed, and sends it.

  • In the inbox your team already uses
  • Grounded in the approved base, never made up
  • Outside the base, a human takes over
See prepared replies
Demo inbox: a customer reports that their domain stopped responding; opposite, a private note from the assistant suggests the reply, visible to the team only.
WHMCS integration

It knows the customer’s real situation

Services, domains, invoices, orders: 0seat reads the customer’s WHMCS account so the reply speaks to their situation, not to a general case.

  • Services, domains, invoices and orders
  • Read-only, through a closed list of actions
  • Password, address and phone never sent to the model
See the WHMCS integration
Voice notes and screenshots

It also understands what the customer doesn’t type

A voice note is transcribed into a private note; an error screenshot is described, so the reply starts from the real problem.

  • Transcription of customer voice notes
  • Screenshots read, never PDFs
  • Two options, off by default, turned on by you
See screenshot reading
Handoff

A human takes over when it matters

Disputes, payments, refunds, cancellations, a request for a person, a second follow-up without a fix: handoff is enforced, whatever the setting.

  • Handoff reason in a private note
  • A suggested message for the agent who takes over
  • No setting can turn it off
See the handoff
Modes per inbox

You decide, inbox by inbox

Every inbox starts off. You switch it to draft for the trial period; automatic replies then take over, category by category, on the inboxes you choose.

  • Off, draft, automatic
  • Payments inbox: draft only, always
  • Automatic on WhatsApp through the official API only
See the modes
0seat console, inboxes: Support in draft mode, Payments in draft mode with the mention “Automatic refused: payments inbox”, Pre-sales in automatic mode.
Costs

Every reply has a visible, capped cost

The cost of every model call is measured and shown in the console, conversation by conversation, under a monthly cap per workspace.

  • Cost measured per conversation
  • End-of-month projection
  • Monthly cap per workspace
See cost tracking
0seat console, Costs tab: current month total, end-of-month projection, bar chart of cost per day over thirty days.

The journey in one minute

A customer message, the reply as a private note, an approved entry, a handoff, the measured cost. Subtitled, no sound.

What comes next

Announced, not yet shipped: these capabilities are not active today.

  • Coming soon

    Real-time voice

    Live voice conversations, with the same approved base and the same handoff rules.

Up and running in three steps

  1. Connect

    An administrator API token is enough: 0seat creates its agent bot in your instance and attaches it to the inboxes you enable. The step-by-step guide lives in the console.

  2. Approve

    0seat extracts candidate entries from your history; your team reviews, edits or rejects them.

  3. Reply

    Drafts arrive as private notes in the inboxes you switch to draft mode. Nothing goes out without an agent.

See how it works in detail

Your data stays yours

One database per space

Each space has its own database: knowledge, conversations, log, costs. Nothing is shared between customers.

Models that retain nothing

Model providers are configured not to retain your data. No free model ever sees customer data without anonymisation.

Encrypted, write-only secrets

Your inbox token and WHMCS credentials are encrypted with AES-256-GCM and never displayed again.

Read everything about security

One price per space, not per agent

Standard plan: $19 excluding tax per space per month, paid in advance, plus model usage at measured cost, capped every month. Add as many agents as you like.

See pricing

Frequently asked questions

Do I need Chatwoot?

Yes. 0seat is an AI layer that runs on Chatwoot, self-hosted or cloud: it connects with an API token and your agents keep their tool. Chatwoot is open source and can be self-hosted. Does your team use another support tool? Write to us: we will look at moving to Chatwoot with you.

Does 0seat answer my customers for me?

Only on the inboxes you choose. Every inbox starts off; in draft mode, the reply arrives as a private note and an agent decides. In autopilot, 0seat answers on its own in the categories where drafts have proven reliable, and hands over everything about money, disputes or cancellations. The server refuses autopilot outright on a payments inbox or an unofficial WhatsApp channel.

What happens when the answer is not in the approved knowledge base?

The conversation goes to a human. 0seat invents no procedure, price or timeline: below the similarity threshold, it does not answer.

Do I have to switch inboxes?

No. 0seat plugs into your existing Chatwoot instance, self-hosted or cloud, through an API token: it creates its agent bot there. Your agents keep their tool.

How much does a draft cost?

The real cost of every model call is measured and shown in the console. It is billed back at your plan’s coefficient, under a monthly cap per space, after which 0seat pauses until the next month.

Is my data used to train a model?

No. Model providers are configured not to retain what is sent, and 0seat trains no model. Each space has its own database.

Judge it on your own history

Create an account, connect an inbox and approve your first entries. Or ask for a demo on your own conversations.