What does time to first audio (TTFA) mean?

It is how long before the first sound of a reply starts to play. For a text-to-speech voice, it is the time from receiving the words to sending back the first sound. For a whole voice agent, it is the time from the caller stopping to the caller hearing the first sound.

What is a good time to first audio?

For a text-to-speech voice alone, published figures run from under 100 ms to about 200 ms. For a whole voice agent, the pause the caller hears should stay under about a second.

What is the difference between time to first audio and time to first word?

Time to first word is the agent deciding and handing its reply to the voice. Time to first audio adds the voice making its first sound, and for a whole agent also speech-to-text and the phone network.

Do bridge phrases like "one moment" count?

They play sound early, so they lower time to first audio without a faster answer. A fair report times the first real answer separately.

Learn · What is time to first audio?

What is time to first audio (TTFA) in voice AI?

Time to first audio is how long before the first sound of a reply starts to play. People use it for two different things, so always check which one a number means.

Komms on live calls: a reply starts in a median of 77 ms

Measured on 1,364 caller turns from 138 real phone calls, October 21 and 22, 2026, from the Komms call logs. The clock runs from the caller's words arriving as text to Komms handing the first words of its reply to the voice.

Komms, live calls	Median	90% within	95% within	Under 200 ms
Reply starts (October 21–22, 2026)	77 ms	123 ms	155 ms	97% of turns
Same line, October 14–20 (6,415 turns)	98 ms	158 ms	177 ms	97% of turns
Same line, early October	162 ms	—	—	—

Real calls, every turn: A live benefits-screening line. No test calls, no lab, and the slowest turns are counted.
Checkable: Every Komms call report shows the timing of every turn, so customers see the same numbers on their own calls.
What it leaves out: Speech-to-text before the clock starts, and the voice and phone line after it stops. Those add time on every platform.

Call it now: +1 (618) 228-1183

Heads up: the demo line is a live Social Security disability (SSDI) intake. It asks real screening questions and calls are recorded. Hang up any time.

Full numbers, method, and what other platforms publish Watch: What is TTFW? (52 seconds)

The two meanings

Meaning	Clock starts	Clock stops	Typical size
For a text-to-speech voice	The voice receives the words	The voice sends back its first sound	Tens of milliseconds to a few hundred
For a whole voice agent	The caller stops talking	The caller hears the first sound	Several hundred milliseconds to over a second

The first is one slice of a reply. The second is the whole pause the caller sits through, the same thing as end-to-end or voice-to-voice latency.

What adds up to the pause a caller hears

Step	What happens
1. Hear	Speech-to-text notices the caller has stopped talking and turns their words into text.
2. Decide	The AI works out what to say.
3. Hand off	The first words of the reply go to the voice.
4. Speak	The voice turns the words into sound, and the phone network carries it to the caller.

Time to first audio for a whole agent is all four steps added together. Time to first word is the agent's part: steps 2 and 3.

What text-to-speech voices publish

Each figure is the vendor's own, read on its own page on October 21, 2026. These cover the voice alone, step 4, not the whole reply.

Voice	Published figure	What it says it measures
ElevenLabs Flash v2.5	~75 ms	Model latency, excluding application and network time. elevenlabs.io
Cartesia Sonic	under 90 ms	First audio. cartesia.ai
Deepgram Aura-2	under 200 ms	Time to first byte. deepgram.com

How to measure it fairly

Say which meaning: A voice-only figure and a whole-reply figure can differ by ten times or more.
Measure from the caller's side: For the whole reply, record real calls and time from the end of the caller's speech to the first sound back.
Separate bridge phrases: A canned "one moment" plays sound early. Time the first real answer too.
Report the median and the slow end: Give the 90th or 95th percentile along with the median, across every turn.

Where Komms fits

Komms publishes time to first word from its call logs, measured on live phone calls (above). It does not yet publish a time-to-first-audio figure, because the moment the first sound plays is not in the call log yet.

Start free Call it now: +1 (618) 228-1183

50,000 free tokens for new Komms accounts — about 625 minutes. No credit card needed.

Questions

What does time to first audio (TTFA) mean?: It is how long before the first sound of a reply starts to play. For a text-to-speech voice, it is the time from receiving the words to sending back the first sound. For a whole voice agent, it is the time from the caller stopping to the caller hearing the first sound.
What is a good time to first audio?: For a text-to-speech voice alone, published figures run from under 100 ms to about 200 ms. For a whole voice agent, the pause the caller hears should stay under about a second.
What is the difference between time to first audio and time to first word?: Time to first word is the agent deciding and handing its reply to the voice. Time to first audio adds the voice making its first sound, and for a whole agent also speech-to-text and the phone network.
Do bridge phrases like "one moment" count?: They play sound early, so they lower time to first audio without a faster answer. A fair report times the first real answer separately.
Does Komms publish time to first audio?: Not yet. Komms publishes time to first word from its call logs, measured on live phone calls, with the method on the Komms latency page.