Telephony & Contact Center
VICIdial
The open-source contact-centre suite, installed and operated properly — by engineers who have run dialers at scale, not just installed the package.
Overview
VICIdial is the leading open-source contact-centre suite. It is built on the Asterisk telephony engine and wraps it in everything a call centre actually needs: a predictive dialer, inbound, outbound and blended call handling, an agent web interface, campaign and list management, call recording, real-time and historical reporting, and the carrier and SIP-trunk integration that connects the whole thing to the outside world. It scales from a handful of seats on one box to many hundreds of agents across a cluster of servers, and it carries no licence cost — that last point is the headline appeal, and the reason it is deployed in call centres across the world.
The catch is the same one that comes with all serious open-source infrastructure: the software is free, and operating it correctly is not. VICIdial is powerful precisely because it exposes the machinery — Asterisk, a MySQL database, the operating system, the dialer’s pacing algorithm, and your relationship with a carrier — and every one of those layers has to be sized, configured and tuned before a busy campaign will hold up. A default install off the ISO will run a demo. It will not survive two hundred agents dialling at a 2.5:1 ratio through a trunk that has not been provisioned for the traffic. That gap between "installed" and "working under load" is where most VICIdial pain lives.
Yarqat works in that gap. We install VICIdial on properly-sized, hardened infrastructure; we tune Asterisk, the database, the carrier configuration and the dialer ratios for the concurrency you actually run; we scale single servers into clustered deployments; we take security seriously, because VICIdial and Asterisk are standing targets for toll fraud; and — very commonly — we are called in to rescue an inherited install that drops calls, loses audio, or falls over the moment the queue fills. We also build the modern touches around it: integrations to your CRM, and AI agents that handle calls alongside your human ones. We operate this software ourselves, so the advice you get is from people who have been woken up by it.
Best for — Call centres that need a capable predictive dialer at scale without per-seat licensing, run by — or with — people who can operate telephony infrastructure properly.
Why teams choose VICIdial
No per-seat licence, at any scale
VICIdial charges nothing per agent. On a floor of two hundred seats that is the difference between a large recurring SaaS bill and hardware you own outright. The saving is real and it grows with headcount — provided the deployment is operated well enough to actually carry those seats.
A full contact centre, not just a dialer
Predictive and manual outbound, inbound queues, blended agents, list and campaign management, recording, and reporting all ship in one integrated suite. You are not stitching together five products; you are configuring one, which is a different and more tractable kind of work.
You own the stack and the data
The servers, the call recordings, the carrier relationships and the database are yours. That means no vendor lock-in, no per-minute markup you cannot see inside, and the freedom to integrate, extend and export exactly as your business needs — which is precisely why serious operators choose it.
Why businesses choose VICIdial
- You run enough agents that per-seat SaaS pricing has become a serious, recurring line item, and owning your dialer would pay for itself.
- You want control over your telephony — carriers, routing, recordings and data — rather than operating inside the limits of a hosted platform.
- You have inherited a VICIdial system that is unreliable under load and need engineers who can actually fix it at the root, not paper over it.
- You want the cost and control benefits of open-source telephony, but you want it installed, hardened and operated by people who have done it before and will carry the pager.
What we build with VICIdial
The capabilities this technology is genuinely strong at — and what we most often build with it.
Predictive dialer and pacing
VICIdial’s predictive dialer places calls ahead of agent availability and connects answered calls to free agents, using a dial ratio and pacing algorithm to keep agents busy without abandoning too many calls. Tuning that ratio against answer rates, average handle time and your abandonment limit is the single biggest lever on a campaign’s efficiency — and the thing most installs get wrong.
Inbound, outbound and blended handling
The same agents can take inbound queue calls and make outbound campaign calls in a blended mode, with VICIdial deciding what to hand them based on queue pressure and campaign settings. Configured well this keeps a floor productive across both directions of traffic; configured badly it starves one to feed the other.
Campaign and list management
Lists are loaded, filtered, recycled and dispositioned per campaign, with dial rules, call-time restrictions, timezones and DNC handling applied to each. This is where regulatory compliance is enforced in practice, and where a large operation’s data hygiene either holds together or descends into chaos.
Call recording and dispositions
Calls are recorded per agent and per lead, with dispositions captured at hang-up to drive reporting and follow-up. Recording carries consent obligations that vary by jurisdiction, and we configure retention and storage deliberately rather than letting recordings silently fill a disk until the system stops.
Real-time and historical reporting
Live agent and campaign screens show what the floor is doing right now — agents on call, in wait, in pause — while historical reports cover calls, dispositions, agent time and campaign performance. These reports run against the same MySQL database that carries live dialing, which is exactly why database sizing and tuning matter under load.
Carrier and SIP trunk integration
VICIdial connects to the outside world through Asterisk SIP trunks to one or more carriers, with dial plans, outbound caller ID and failover routing configured per your traffic. Provisioning enough concurrent channels with a carrier, and configuring the trunk correctly, is what separates a system that dials cleanly from one that drops calls and loses audio at peak.
Use cases
Outbound campaign call centres
Sales, collections, surveys and appointment-setting operations running predictive outbound to large lists, where dialer efficiency directly drives revenue and where per-seat licensing at scale would be prohibitive.
Blended inbound and outbound floors
Operations where agents both take inbound enquiries and make outbound calls, needing one system that balances both intelligently across a shared pool of agents rather than two disconnected tools.
Rescuing a failing inherited install
A VICIdial system — often set up by someone long gone — that drops calls, delivers one-way audio, or collapses when the queue fills. We diagnose across Asterisk, the database, the OS and the carrier, and stabilise it rather than starting from scratch.
Adding modern capability around VICIdial
Integrating an existing dialer with a CRM over its API, or putting AI voice agents on the front line to qualify or handle calls alongside human agents, so an established VICIdial floor gains modern capability without being ripped out.
When VICIdial is the right choice
- Right when you are running a real call centre and need a serious, capable dialer — predictive outbound, blended inbound, campaign and list management — at a scale where per-seat SaaS licensing would cost you a fortune every month. VICIdial’s zero licence cost is genuine, and at scale it is decisive.
- Right when you have the appetite to operate telephony infrastructure, either with your own people or with us alongside you. VICIdial rewards teams who want control over their dialer, their carrier relationships and their data, rather than accepting whatever a hosted vendor allows.
- Right when you have inherited a VICIdial install that misbehaves under load and you need someone who can diagnose it at the Asterisk, database and carrier level rather than reinstalling and hoping — rescue work is some of the most valuable we do here.
- Wrong for a tiny team of a few agents with no interest in telephony. A hosted SaaS dialer will get you running in an afternoon with no servers to patch, and for low seat counts the licence cost is trivial against the operational burden VICIdial adds. We will tell you so.
- Wrong for any team with zero appetite for operating infrastructure. VICIdial is not fit-and-forget. Somebody owns the servers, the carrier, the security posture and the regulatory compliance. If nobody on your side wants that responsibility and you do not want to pay us to carry it, hosted is the honest answer.
VICIdial: pros and cons
Strengths
- Zero licence cost, which becomes a decisive advantage as agent counts and call volumes grow.
- A complete, integrated contact-centre feature set — predictive dialing, blended inbound/outbound, recording and reporting — proven in production across thousands of real call centres.
- Built on Asterisk and open standards, so it integrates with SIP carriers and your own systems rather than trapping you in a closed platform.
- Scales genuinely, from one server to clustered multi-server deployments serving hundreds of concurrent agents, when it is architected and tuned for it.
Trade-offs
- It demands genuine expert installation and operation. A default install will not hold up at scale — Asterisk, the database, the OS and the carrier all need tuning for your real concurrency, and getting that wrong shows up as dropped calls and one-way audio.
- It is operationally demanding. You own the servers, the patching, the carrier relationship and the monitoring; telephony problems can involve any of those layers, and diagnosing them needs breadth most teams do not have in-house.
- The agent and admin interface is dated compared with modern commercial SaaS dialers. It is functional and dense rather than polished, and agents used to slick tools will notice.
- You own the compliance and security burden entirely — telemarketing regulations, call-recording consent, and toll-fraud risk. VICIdial and Asterisk are actively targeted by attackers, and an unhardened box can run up enormous fraudulent call charges overnight.
Architecture
A VICIdial deployment is several cooperating parts, and treating it as one black box is the first mistake. At its centre is Asterisk, the telephony engine that originates and bridges every call. Around it sit a web/agent layer (Apache and PHP serving the agent and admin interfaces), a MySQL database that holds leads, campaigns, live agent state and every call record, and the carrier connection over SIP that carries real audio in and out. On a small system these can share one server; at scale they must not.
We architect for the concurrency you actually run. That means separating the database onto its own well-provisioned host once call volume justifies it, running multiple dialer/telephony servers behind the setup so no single Asterisk box is asked to bridge more channels than it can, and keeping the web layer able to serve agent sessions without contending with dialing for resources. The clustered VICIdial topology — one database, multiple dialers, a load-balanced web tier — is a well-trodden pattern, but it only works when the pieces are sized against measured load rather than guessed.
The carrier relationship is part of the architecture, not an afterthought. The number of concurrent channels you have provisioned with your SIP carrier is a hard ceiling on how many calls you can place, and outbound routing, caller ID and failover all live in the Asterisk dial plan. We design that layer deliberately, so the system degrades sensibly under pressure instead of dropping calls silently.
Performance
VICIdial performance is really the performance of four things at once: Asterisk’s ability to bridge concurrent channels, the database’s ability to serve live dialing and reporting queries without stalling, the operating system’s handling of interrupts and real-time audio, and the carrier’s ability to carry the offered traffic. A slow campaign, choppy audio or agents left waiting is a symptom, and the fix depends entirely on which layer is the bottleneck.
We tune with measurement, not folklore. That means watching Asterisk channel counts and CPU under real dial pressure, profiling the MySQL queries that back live and historical reporting and indexing or separating them so reporting never competes with dialing, and checking the operating system and network stack for the latency and jitter that ruin call audio. The dialer’s pacing ratio is then tuned against actual answer rates and handle times — pushed hard enough to keep agents busy, held back enough to stay inside your abandonment limit.
Honesty matters here: there is no single setting that makes VICIdial fast. Performance comes from sizing each layer for the concurrency you run and keeping the layers from starving one another. That is ongoing work, not a one-time switch, which is exactly why operating experience beats a clever initial install.
Security
This is the part too many VICIdial deployments treat as optional, and it is the part that costs people the most. VICIdial and Asterisk are standing targets for toll fraud: attackers scan the internet for exposed SIP and admin interfaces, brute-force weak credentials, and route expensive international calls through a compromised box. An unhardened server can run up thousands of pounds in fraudulent charges overnight, and the carrier will still expect to be paid.
We treat hardening as mandatory, not a nice-to-have. That means the SIP and management surfaces are not open to the world; strong, non-default credentials everywhere; firewalling and rate-limiting on SIP registration and Asterisk Manager access; fail2ban or equivalent watching for brute-force attempts; carrier-side controls such as international-dialing restrictions and spend alerts; the OS kept patched; and monitoring that tells you when call patterns look wrong before the bill does. Recording and data handling are secured deliberately too, because call recordings are sensitive personal data.
None of this is exotic, but all of it has to be done and then maintained. Security on a VICIdial box is not a state you reach once; it is a posture you hold, and it is one of the clearest reasons to have people who operate these systems looking after yours.
Scalability
VICIdial genuinely scales — from a single server running a handful of agents to a cluster serving many hundreds of concurrent seats — and that is one of its real strengths over lighter tools. But scale is not something you flip on; it is an architecture you grow into, and the move from one box to many is where installs most often break if they were never built for it.
The path is well understood. First the database comes off the shared server onto its own tuned host, because at scale it is the component under the most contention from both live dialing and reporting. Then telephony load spreads across multiple Asterisk/dialer servers so no single box exceeds the channels it can bridge cleanly. In parallel, the carrier side scales — more provisioned concurrent channels, sensible outbound routing across trunks — because your dial capacity is capped by what the carrier will carry regardless of how many servers you add.
We plan the growth ahead of the need, so adding a dialer server or database capacity is a routine step rather than an emergency. The floors that fall over at two hundred agents are almost always the ones that were sized for twenty and never re-architected. We size for where you are going, not just where you are.
VICIdial integrations & ecosystem
The technologies we most often pair with it — each links to how we work with it.
How we work
We start by understanding the real operation: how many agents, what campaigns, inbound, outbound or blended, what call volume at peak, which carrier, and — if it is a rescue — exactly how the current system fails and when. That shapes everything, because a twenty-seat outbound campaign and a three-hundred-seat blended floor are different engineering problems that happen to share software.
From there we install on properly-sized, hardened infrastructure — or, for a rescue, we diagnose across Asterisk, the database, the OS and the carrier before we touch anything, so we fix the cause rather than the symptom. We tune the dialer ratios, the database and the carrier configuration against measured load, harden the system against toll fraud as a matter of course, and set up monitoring so problems announce themselves early. Where you want modern capability around it — CRM integration over the API, or AI voice agents on the floor — we build that as a deliberate extension, not a bolt-on.
The engineers who set your system up are the ones who understand it end to end, and we are candid throughout about trade-offs — including telling you when hosted SaaS would genuinely serve you better than owning this. We operate telephony ourselves, so we build for the two-in-the-morning reality, not the demo.
The service behind it
Delivered throughCall Center SetupWhat we build with VICIdial
The disciplines this technology most often shows up in — from a first build to taking over and stabilising an existing one.
How we deliver
- 01
Discover
We map the system, the constraints and the business it serves — including the parts nobody documented.
Architecture brief
- 02
Architect
Decisions get made, written down and defended before a line of production code exists.
Decision records
- 03
Build
Short cycles against working software. You see progress in the product, not in a status deck.
Shipping increments
- 04
Operate
Monitoring, incident response and iteration. The system is alive, so the engagement is too.
Runbooks & SLOs
Industries we use VICIdial in
Domain knowledge changes what gets built. A few of the sectors we know before the first meeting.
Also in Telephony & Contact Center
Why teams choose us for VICIdial
We operate telephony, not just install it
We run VICIdial and Asterisk systems ourselves, so we build for the reality of a busy floor and a two-in-the-morning failure — not for a clean demo. The person configuring your dialer knows what breaks under load because they have been on the receiving end of it.
Rescue is a core competence
Diagnosing a system that drops calls or loses audio needs someone who can reason across Asterisk, MySQL, the OS and the carrier at once. That breadth is exactly what most inherited VICIdial installs lack, and exactly what we bring.
We take security seriously by default
Toll fraud is not hypothetical — it is a nightly threat to any exposed VICIdial box. We harden as standard and treat compliance and consent as real obligations, because we have seen what an unhardened system costs when nobody was watching.
Senior engineers, honest advice
No juniors learning telephony on your call centre. And if your seat count means a hosted SaaS dialer would genuinely serve you better than owning this, we will say so before you spend the money.
Typical timeline
- 01
Discovery and sizing
One to two weeks understanding the operation — seats, campaigns, peak volume, carrier — or, for a rescue, diagnosing exactly how and when the current system fails, so the infrastructure is sized against reality before anything is built.
- 02
Install, harden and configure
Building the deployment on properly-sized infrastructure, configuring campaigns, carrier and recording, and hardening against toll fraud from the outset — security is part of the build, not a later phase.
- 03
Load tuning and validation
Tuning dialer ratios, the database and the carrier against real dial pressure, and validating call quality and agent experience under concurrency close to your peak before the floor goes live.
- 04
Go-live and ongoing operation
Cutting the floor over with monitoring in place, then operating and scaling the system — patching, security, tuning and adding capacity — as the operation grows.
How pricing works
- Fixed-scope installation and hardening for a defined deployment — a specific seat count, campaign type and carrier — quoted once we understand the operation and its concurrency.
- Rescue and stabilisation engagements for a failing inherited system, scoped from an initial diagnostic across Asterisk, the database, the OS and the carrier.
- Ongoing operation and support — patching, security, monitoring, tuning and scaling — for teams that want us to carry the operational burden rather than staff it themselves.
- Focused extension work — CRM and API integration, or AI voice agents around an existing VICIdial floor — priced by the defined scope.
Hire VICIdial engineers
Need VICIdial capacity on your own team? We embed named senior engineers into your existing team — reporting to your leads, working in your rituals — so you add capacity without a hiring cycle.
Hire VICIdial engineersCommon questions
Is VICIdial really free?
The software carries no licence cost, and that is genuine — you pay nothing per agent, at any scale. What is not free is running it: servers, carrier minutes and channels, and the expertise to install, harden, tune and operate it. On a large floor the licence saving is decisive against SaaS; on a tiny team the operational cost can outweigh it. We are honest about which side of that line you sit on.
Why does our VICIdial drop calls or lose audio under load?
Almost always because one of the layers is under-provisioned or misconfigured for your real concurrency — too few carrier channels, an overloaded Asterisk box, a database stalling on reporting queries during live dialing, or OS/network latency ruining audio. It is a diagnosis job across all of those, not a single setting. Reinstalling rarely fixes it because the cause is architectural. This kind of rescue is some of the most common work we do.
How many agents can VICIdial handle?
From a handful on a single server to many hundreds across a clustered deployment. The number depends entirely on architecture — one box has a real ceiling on concurrent Asterisk channels and database load. Scaling past it means separating the database, spreading telephony across multiple dialer servers, and provisioning enough carrier channels. It scales genuinely, but only when it is built and tuned for the scale you are actually running.
How worried should we be about security and toll fraud?
Very. VICIdial and Asterisk are actively scanned and targeted, and an exposed, unhardened box can be hijacked to route expensive international calls, running up thousands in fraudulent charges overnight — which the carrier will still bill you for. Hardening is mandatory: locked-down SIP and admin surfaces, strong credentials, firewalling, brute-force protection, carrier-side dialing limits and spend alerts, and monitoring. We do this as standard, and we would urge anyone running VICIdial to treat it as non-negotiable.
Can we add AI voice agents or connect VICIdial to our CRM?
Yes. VICIdial exposes an API and sits on open standards, so we integrate it with a CRM to sync leads, dispositions and recordings, and we can put AI voice agents on the floor to qualify or handle calls alongside human agents. We build these as deliberate extensions around a properly-operated system — the dialer and its infrastructure have to be solid first, or you are adding modern capability on top of an unstable base.
Building on VICIdial?
A technical conversation with the engineers who would do the work. If we are not the right fit, we will say so on the call.