For contact centres, BPOs and remote workforce teams
Prove an agent's connection is fit for the shift
A browser test your work-from-home agents run on their own line before they take calls. Download, upload, latency and jitter go to your platform and your dashboard, per agent, per shift.
30-day free trial · 100 API calls · No credit card · Plans from $49/month
This is the check an agent would run. In your portal the numbers go to your code and your dashboard, keyed to the agent.
Put it in the agent portal
One script tag on the shift-start page, or a link a supervisor sends. Agents need no software beyond the browser.
Compare with your softphone's needs
Latency and jitter decide call quality; upload decides whether screen share and video hold. Your callback gets all four numbers.
Clear the agent, or route them
Pass them into the queue, tell them what to fix, or open a ticket with the numbers attached. Every result is in your dashboard by date.
The connection is the part you do not control
A remote agent's desk is a home broadband line and a Wi-Fi router you have never seen. When calls drop or audio turns robotic, the softphone, the VPN and the agent all get blamed before the line does, and by then the customer has hung up. Onboarding checks catch the worst cases once; they say nothing about the line on a Monday morning in the rain.
A readiness check at shift start measures the line the agent is actually on, in the browser they are already using, and gives your workforce team a number instead of an argument. One customer named the subdomain it runs on for exactly this job: wfhtest.
What to measure for voice
| Metric | What it does to a call | Typical requirement |
|---|---|---|
| Latency | Delay on the line; agent and customer talk over each other | Under 150 ms |
| Jitter | Robotic or choppy audio on a line that looks fast | Under 30 ms |
| Upload | The agent's voice, plus video and screen share where used | 1 Mbps and up; more for video |
| Download | Customer audio, CRM and knowledge tools | Rarely the limit on home broadband |
Voice needs little bandwidth and a lot of stability, which is why a download-only check misses most problems. The test measures a single stream of real files against the nearest of our servers, so the latency and jitter figures reflect the route the agent's calls will take out of the home, not an idealised lab number.
What your platform receives
| Property | Type | What it means |
|---|---|---|
| download | number | Average download speed in Mbps |
| upload | number | Average upload speed in Mbps |
| maxDownload, maxUpload | number | Peak speeds in Mbps |
| latency | number | Minimum round trip in milliseconds |
| jitter | number | Latency variance in milliseconds |
| testServer | string | Which test server the browser reached |
| ip_address, hostname, userAgent | string | Who ran it, for your records |
The result arrives in your page, so it can be keyed to the agent, posted to your workforce management or ticketing system, and used to gate the queue. Every test is also logged in your analytics dashboard with its date, location and device, so a supervisor can look up an agent's history without asking IT.
Built for the scale of a floor
The check runs from a script tag, so it scales from a pilot team to every agent on every shift without infrastructure on your side; the plans go from 1,000 checks a month to the millions the largest floors have run through it. Support teams use the same integration the other way round: when an agent reports call problems mid-shift, the supervisor sends the link and reads the numbers.
Hiring platforms, contact centres and ISPs have run their connection checks on this API since 2013.
What it costs
Every completed test is one API call. The 30-day trial includes 100 calls and needs no card. Paid plans start at $49 a month for 1,000 calls (Startup), then $149 for 10,000 (Business) and $599 for 100,000 (Pro); enterprise volume is priced on request. The full table is on the API page.
Questions people ask
Can agents run it on their own without IT?
Yes. It runs in any modern browser on any device, with nothing to install. A link in the agent portal or in a supervisor's message is enough.
Can we run it automatically every shift?
Yes. Your page decides when to call startTest(), so it can run on login, on the shift-start screen, or on a schedule you set, and gate the queue on the result.
Does it work through our VPN or VDI?
It measures whatever path the browser takes. Running it inside the VPN or virtual desktop measures the route calls will actually use; running it outside measures the raw home line. Many teams check both when they set thresholds.
Where do the results go?
To your callback in the page, and to your analytics dashboard with date, location and device. Nothing needs to be built to see the first results; a database only comes in when you want to join them to your own agent records.
Can we see an agent's connection over time?
Yes. Every test lands in your dashboard with its date, and the callback in your portal can write the same numbers against the agent's record, so a supervisor can see whether Monday's dropped calls line up with a line that has been degrading for a week or with one bad morning.
30-day free trial · 100 API calls · No credit card · Plans from $49/month