Telnyx AMD pricing: standard vs premium, and when it applies

Telnyx AMD pricing at a glance

Telnyx offers answering machine detection in two tiers. According to Telnyx's public release notes as of November 2026:

• Standard AMD: $0.0020 per invocation, or $2.00 per 1,000.
• Premium AMD: $0.0065 per invocation, or $6.50 per 1,000.

Both prices are on top of call charges. You pay for the minutes and you pay again each time detection runs. Telnyx can change prices at any time, so check its current pricing page before you plan around these numbers.

The second point is easy to miss: Telnyx AMD only works on calls placed and controlled through Telnyx's API. If your VICIdial or Asterisk server dials out over a Telnyx SIP trunk, Telnyx is carrying the audio but is not running the call, and its AMD never sees it. Many call centers find this out after signing up for the trunk.

Standard vs premium: what the price difference means

Premium AMD costs 3.25 times as much as standard per invocation. Telnyx's own documentation describes the difference between the tiers, and it is worth reading before you choose. From a buyer's point of view, the question to answer is simple: does the extra $4.50 per 1,000 calls buy you enough extra live conversations to pay for itself?

You can only answer that with your own calls. A sensible test looks like this:

1. Pick one campaign and split its traffic between the two tiers for a week.
2. Pull a random sample of calls each tier marked as a machine, and have a person listen to the start of each.
3. Count the live people each tier sent away, and the machines each tier sent to agents.
4. Multiply the difference in live people by what a conversation is worth to you, and compare that to the price gap.

The same method works for any detection product. Our post on what AMD false positives cost walks through the math, and the ROI calculator does it for you in the browser.

How Telnyx AMD works with Call Control

Telnyx answering machine detection is part of Telnyx Call Control, its API for placing and steering calls from your own application.

• Turning it on: when your application dials through Call Control, it includes the answering_machine_detection option on the request. The exact values for each tier and mode are in Telnyx's API reference.
• Getting the result: detection results arrive as webhook events sent to your application.
• Acting on it: your application reads the event and sends the next command: connect to an agent, play a message, or hang up.

This model fits well if you are building your own calling application on Telnyx. You get detection as one more event in a flow you already control. It does mean your application, not a dialer like VICIdial, is in charge of the call. You need a public endpoint for webhooks, code that handles each result, and logic for timeouts and unexpected events.

Costs beyond the per-invocation fee

• Development and upkeep. Webhook handling, retries and routing rules are code your team writes and maintains. That time has a cost even if it never shows up on an invoice.
• Dead air. Every moment spent deciding, then waiting for a webhook and sending the next command, is silence on the line for a live person. Long silences lead to hang-ups and can push abandoned calls up; our dead air guide explains why.
• Wrong answers. Each invocation costs the same whether the result is right or wrong. A live person sent away as a machine is a paid-for lost conversation.

Why Telnyx AMD doesn't apply to Asterisk over a Telnyx SIP trunk

Most VICIdial and Asterisk centers that use Telnyx use it as a SIP trunk. In that setup the call flow looks like this:

1. VICIdial's hopper hands a lead to Asterisk.
2. Asterisk sends a SIP INVITE out over the Telnyx trunk.
3. The far end answers, and audio flows back to your Asterisk server.
4. Asterisk runs your dialplan. With AMD on, VICIdial sends the call to extension 8369, which runs Asterisk's AMD() and then VD_amd.agi to route on the result.

Telnyx never receives a Call Control command for that call, so there is nothing to attach the answering_machine_detection option to. The detection decision happens inside your own dialplan, on your own server. To use Telnyx AMD you would have to move call control out of Asterisk and into an application built on Telnyx's API. For a VICIdial center that means giving up the dialer's own call handling, which is rarely worth it just to change AMD.

The practical answer is to keep the trunk and run better detection where the call already is: on the Asterisk server. Our Telnyx AMD alternative page compares the two approaches in more detail.

Telnyx AMD cost examples

The table uses the November 2026 per-invocation prices and assumes one invocation per answered call. It shows detection fees only. Call minutes are extra in every column, including nomoVM's, since nomoVM doesn't carry your calls.

Two things to check against your own setup:

• What counts as an invocation. Telnyx prices AMD per invocation. If your application requests detection on calls that are never answered, ask Telnyx whether those are charged, and adjust the table.
• Your negotiated rate. High-volume accounts may have different pricing. Use your contract rate if you have one.

For a side-by-side view that adds Twilio and Asterisk's built-in AMD() at 1, 3 and 10 million calls, see our answering machine detection cost comparison. For Twilio specifically, read Twilio AMD pricing.

Keep your Telnyx trunk, add nomoVM

nomoVM is hosted AI answering machine detection for VICIdial, GoAutoDial, Asterisk and FreeSWITCH. Your dialer streams the first seconds of each answered call to wss://amd.nomovm.com/ over an encrypted WebSocket and gets one verdict back. Because it listens to audio your dialer already has, it works with Telnyx and any other SIP carrier. Your trunk, routing and agents don't change.

There is no additional infrastructure required. No new servers, no hardware, no carrier change: one script on the dialer you already run. Setup takes about 10 minutes.

On Asterisk or VICIdial

Copy nomovm.agi.py to /var/lib/asterisk/agi-bin/, make it executable with chmod 755, and test it:

python3 nomovm.agi.py --test wss://amd.nomovm.com/ YOUR_KEY

It prints AMDSTATUS=MACHINE AMDCAUSE=SILENCE in about 2 seconds. Then add one line to your AMD extension:

exten => 8370,n,EAGI(nomovm.agi.py,wss://amd.nomovm.com/,YOUR_KEY,4500)
exten => 8370,n,GotoIf($["${AMDSTATUS}" = "FALLBACK"]?native)

On VICIdial you copy the 8369 extension to 8370 with these lines added and switch one campaign's Routing Extension to 8370. Setting it back to 8369 rolls back. The complete dialplan is in the VICIdial AMD setup guide, and plain Asterisk is covered on the Asterisk AMD page.

What happens on each call

• Most calls are decided about 2 seconds after answer; unclear calls listen up to 3.2 seconds.
• AMDSTATUS is HUMAN, MACHINE or FALLBACK. AMDCAUSE gives the reason, including SCREENING for Google Call Screen and iPhone screening.
• If an error or timeout happens after streaming starts, the call is treated as MACHINE/UNKNOWN, so only a confirmed person reaches an agent.
• If the service can't be used at all, FALLBACK comes back before any audio is used, Asterisk's own AMD() runs, and the call is never billed.

Pricing

Free covers 30,000 calls a month with no card. Starter is $99 for 1,000,000 calls, Growth $199 for 3,000,000, Scale $499 for 10,000,000, and extra calls on paid plans are $0.0002 each. There are no per-server or per-agent fees and no limit on servers. See the pricing page for details, and the accuracy page for both of our measured error numbers.

Next steps

If you are building a calling app on Telnyx Call Control, test standard and premium AMD on your own traffic and price them with the table above. If your Asterisk or VICIdial server dials over a Telnyx trunk, keep the trunk and put detection on the dialer: start a free nomoVM account and follow the setup guide on one campaign.

Frequently asked questions

How much does Telnyx AMD cost?

According to Telnyx's public release notes in November 2026, standard AMD costs $0.0020 per invocation and premium AMD costs $0.0065 per invocation. Both are charged on top of call charges. Check Telnyx's current pricing page before budgeting.

What is the difference between Telnyx standard and premium AMD?

They are two detection tiers at different prices: $2.00 versus $6.50 per 1,000 invocations. Telnyx's documentation describes what each tier does. The practical way to choose is to test both on your own calls and compare how many live people each sends to your agents.

Does Telnyx AMD work if VICIdial uses Telnyx as a SIP trunk?

No. Telnyx AMD is an option on calls you dial and control through Telnyx Call Control. When your own Asterisk or VICIdial server places calls over a Telnyx SIP trunk, Telnyx only carries the audio, so its AMD is not part of that call. Detection has to run on your dialer instead.

How do I get Telnyx AMD results?

You enable the answering_machine_detection option when dialing through Telnyx Call Control, and the result arrives as a webhook event. Your application receives that event and decides whether to connect the call, leave a message or hang up.

Can I keep Telnyx as my carrier and use nomoVM?

Yes. nomoVM listens to audio your dialer already has, so it works with any SIP carrier, including Telnyx. Your trunk, routing and agents don't change; you add one EAGI line to your Asterisk dialplan.

What does nomoVM cost compared with Telnyx AMD?

nomoVM plans are $99 a month for 1,000,000 calls, $199 for 3,000,000 and $499 for 10,000,000, with $0.0002 per extra call. At 1 million answered calls, Telnyx AMD fees would be $2,000 standard or $6,500 premium. Carrier minutes are paid separately in every case.