Skip to content
EN

Deployment options

The platform on your hardware

You can install CONTACT.WORK in your own server room. Conversations and their history then stay inside your company, and the platform can pull data from systems that have no internet access. Your existing Asterisk telephony stays as is – the platform connects to it. It takes two to seven days from order to go-live. There's no server rental to pay for: you pay only for installation and a support plan.

2-7 daysto go livesingle server or cluster
$0hosting and license feesthe hardware is yours
1 serverfor 10 agentsruns the whole platform
4 serversfor 200 agentseach with its own role

Capabilities

What you get with on-premises

Data stays in-houseCall recordings, chats, and customer forms never leave your company.
Closed systemsAccounting, inventory, and customer databases connect to the platform inside your office network, with no outside access.
Works offlineIf the internet goes down, customer records, reports, internal calls, and a separate phone line keep working in the office.
Your own telephonyYour existing phone system and numbers stay – the platform fits into your setup.
AI with no per-minute feesAI models run on your side; one of the servers gets a GPU for them.
No hosting or license feesYou pay for support and installation. We don't charge for the platform or for seats.
Your access rulesYour security team decides how employees and contractors get into the system.
MonitoringWe set up Zabbix and Grafana monitoring, so you know about resource shortages in advance.

Use cases

When companies choose their own servers

Running it yourself makes sense when keeping data in-house and connecting internal systems matter more than a fast start.

Storage required by law

A bank, a clinic, or a government agency can't let call recordings and customer forms leave the building – by law or by security policy. We install the platform in your server room, and nothing goes outside it. Responsibility for the personal data stays with you. We hand over the platform architecture and a description of which data goes where for review before launch.

  • Everything on-site
  • You stay in control
  • Documents up front
Example 1 of 4

If some of your data can live in the cloud, take a look at hybrid deployment.

Architecture

How the installation is set up

Servers

How many servers you need

A small team needs just one server: a 4-core CPU, a 240 GB SSD, and 16 GB of RAM. With more agents, platform components are split up: telephony, applications with the database, service modules, and reports each get their own server. Both physical and virtual Linux servers work.

  • 1 to 4 servers
  • Linux
  • Physical or virtual

Recordings

Where calls are stored

Call recordings get a dedicated hard drive – 1,000 hours of calls take up about 90 GB. Along with the combined recording, the platform puts each speaker's voice on a separate track; without that, speech analytics doesn't work. You decide how long to keep recordings.

  • SATA drive
  • Two tracks
  • Your retention period

Redundancy

Cluster and monitoring

If downtime is expensive, we add standby servers to the main ones and combine them into a cluster: if one fails, another takes over the load. Monitoring warns you in advance when disk space is running out, memory is low, or a connection drops. We also handle backups, and they're stored on your side.

  • Cluster
  • Backups on your side
  • Early warnings
Nothing leaves your networkCall recordings and customer forms are stored only inside your company.
Internal systems at handData from accounting systems and databases with no internet access opens right in the customer record.
Keep your telephonyAsterisk, phone lines, and numbers work as before.
No server rentalThe hardware is yours – you pay for support and installation.

Installation

How we install the platform

  1. 01AssessmentWe review your network, telephony, and internal systems, and estimate the number of agents and disk space you'll need for the years ahead.
  2. 02RequirementsWe give your IT team a server list and network parameters, and your security team a description of the platform and how data flows.
  3. 03PreparationYour IT team or ours prepares the servers: we install Linux and set up the network, users, and access.
  4. 04InstallationWe install the platform and telephony, move your numbers to the new setup, and connect your CRM and business systems.
  5. 05TestingWe test calls, call recording, and reports with a few agents, and turn on server monitoring and alerts.
  6. 06Go-liveWe roll the platform out to all agents, train supervisors, and support the platform under a support contract.

Comparison

Licensed software vs. CONTACT.WORK

CONTACT.WORKLicensed software
Seat feesNoneA license for every agent
ModulesAll includedBought separately
New agentsFree while your servers have capacityNew licenses
UpdatesIncluded in supportOften paid
CustomizationPaid from support hoursWaits in the vendor's queue
SupportOne teamVendor plus integrator

Requirements

What you'll need

Small team
One server
Up to 200 agents
Four servers
Operating system
Linux: Rocky Linux
Recordings
A dedicated recording drive
On-site AI
A GPU
Price
$0 for the platform, work billed in plan hours

Before you start

What to decide up front

Who will maintain it?Your admins, CONTACT.WORK engineers under a support contract, or both together.
How many agents will you have?A three-year forecast shows how many servers to buy now and how many later.
Do you need redundancy?If an hour of downtime is expensive, plan a cluster and standby servers from the start.
How long will you keep recordings?Six months, three years, or longer – the retention period sets the size of the recording drive.
Which systems should we connect?Which internal systems need to be connected and where call outcomes should be saved.
Do you need AI without the cloud?Then you'll need a GPU installed in your least loaded reporting server.

Q&A

Frequently asked questions

How long does installation take?

Two to seven days. A single server is faster; a redundant cluster takes longer.

What servers will we need?

One for 10 agents, two for 30, three for 60, and four for 200. The specs for one server are listed above; we size the exact configuration for your project.

How much does on-premises cost?

The platform is free, and there's no server rental. You pay for a support plan (the simplest is $180 a month), installation work from plan hours, and your own hardware – servers and drives. You pay your carrier for call minutes.

Do we need internet access?

If the hardware is in your office, the platform works on the office network without internet. Internet is needed for messaging channels and cloud AI models, if you use them. External calls keep working too, as long as your phone line doesn't come from the same provider as your internet.

Where can we put the servers?

Wherever suits you: an office server room, your own data center, or any colocation facility you choose.

How do remote agents work?

They connect to the company network through the office VPN, and from there it's just like working at an office desk.

Who makes backups?

We do. Backups are kept on your infrastructure, and our support covers configuring and verifying them.

Can we migrate from our old system?

You can. We load your existing data into the platform ourselves – import and export are included in support plans from Standard up.

Who is responsible for the personal data?

Your company. The data is on your side, and you decide who has access to it.

How do your engineers get to our servers?

Through access your admins grant. Every connection is recorded in the audit log.

Who updates the platform?

We do, on a schedule we agree on with you. Updates are included in every support plan.

Will AI agents work without the cloud?

They will if your server room has a GPU. It goes into the least loaded reporting server. There are no per-minute fees then, and agent setup and training come out of your support hours.

Can we run speech analytics on-site?

You can. The platform records the agent's and the customer's speech on separate tracks, and transcription runs on your side, on a server with a GPU.

Will we keep our telephony?

You will. Asterisk and your carrier lines keep working; the platform simply connects to them. Your numbers stay the same.

Do you monitor the servers?

We do: we combine servers into a cluster so work continues on another server if one fails, and set up monitoring with alerts so engineers know about an outage right away.

Do virtual servers work?

They do. The platform runs on both physical and virtual Linux servers.

Can we start in the cloud and move in-house later?

You can. The platform moves with its call flows, integrations, and settings, and we switch your numbers over at a scheduled time.

Glossary

Terms on this page

Server cluster
A group of machines that share the load and back each other up. Learn more →
Cloud deployment (cloud)
The platform runs on the vendor's servers, and agents connect over the internet. Learn more →
Call recording
Calls are saved so they can be replayed and disputes settled. Learn more →
Backup
Data is regularly copied to a separate place so it can be restored if something goes wrong. Learn more →
Customer record
The screen of customer data an agent sees during a conversation. Learn more →
Hosting
You rent a server from a provider instead of buying one. Learn more →
Speech analytics
The system analyzes recorded calls automatically: topics, script adherence, customer emotions. Learn more →
On-premises deployment
The system runs on the company's own servers, not in the vendor's cloud. Learn more →
Personal data
Anything that points to a specific person: a name, a number, a voice. Learn more →
SIP (Session Initiation Protocol)
The protocol that sets up calls in IP telephony. Learn more →
Hybrid deployment
Part of the system runs on the company's servers, part in the vendor's cloud. Learn more →
CRM – customer relationship management
Where a company keeps its customers, deals, and full communication history. Learn more →
Supervisor
Leads a team of agents: watches the queue and quality, helps and coaches. Learn more →
Data center
A building with servers, backup power, cooling, and network links. Learn more →
VPN
An encrypted tunnel from the internet into an internal network. Learn more →
Dual-channel recording
Each speaker is recorded on a separate track. Learn more →
Virtual server
A software-defined machine that shares hardware with others. Learn more →

Blog

From the blog

All articles →

Updated: