Asterisk AMD,
without the stopwatch.

Add acoustic AI detection to any Asterisk 13+ dialplan with a single EAGI step. Your carrier, routing and agents stay exactly as they are.

Start free See all integrations

125ms

detection starts after answer

calls lost when AMDY is unreachable

50K

free detections to evaluate

THE LIMITS OF AMD()

Timing can’t hear
the difference.

AMD() counts silences, words and their lengths. On modern compressed call audio, those signals overlap between people and recordings.

NOTSURE eats the window

Unclear calls use the full analysis time, so the person hears silence and hangs up.

Noisy lines look like machines

Background sound and clipped words confuse timing rules.

Recorded “Hello?” robots pass

Recordings built to sound human have human timing, too.

HOW IT FITS

Add a step.
Keep AMD() as backup.

The EAGI script reads the call’s audio from Asterisk, streams it to AMDY and sets AMDSTATUS. If anything goes wrong, it returns FALLBACK before any audio is used and your existing AMD() line takes the call.

Read the API docs

EXTENSIONS.CONFEXAMPLE

exten => 8370,n,EAGI(amdy.agi,KEY)
exten => 8370,n,GotoIf($["${AMDSTATUS}" = "FALLBACK"]?native)
exten => 8370,n(vdamd),AGI(VD_amd.agi,${EXTEN})
exten => 8370,n,Hangup()
exten => 8370,n(native),AMD()
exten => 8370,n,Goto(vdamd)

FALLBACK KEEPS NATIVE AMD8 KHZ · MONO

WHAT CHANGES

For your agents,
for your numbers.

The dialer still dials and your carrier still carries the call. The decision just gets better.

Any dialplan

Works with custom Asterisk dialers, FreePBX and Issabel builds.

Any carrier

AMDY hears the audio your server already has. No trunk changes.

Fail-safe by design

A wrong key or network problem never costs you a call.

READY FOR A BETTER FIRST HELLO?

Try AMDY on Asterisk.

50,000 free detections. Point one campaign at it and compare.

Talk to AMDY

Timing can’t hearthe difference.