If you've worked with expressive TTS in the last year you've probably seen the pattern:
She d. [sigh] "Fine, you can come in."
Inline paralinguistic tags. Half the model demos use them. So when we wired up Inworld TTS-1.5 Max for HoneyChat — Telegram-native AI companion where voice messages are a first-class output — we sprinkled [laugh]
, [sigh]
, [breathe]
through the prompts and shipped.
The audio sounded fine. Just… exactly the same as before. No laugh. No sigh. The tags were getting read out as silence at best, and as the literal text "sigh" at worst, depending on the voice.
We tested all the variants we could find. None of them moved the needle.
HoneyChat voice stack at a glance:
en, ru, ja, zh, ko, es, fr, de, it, pt, pl, hi, ar, he, nl
.voiceId
strings in config/archetype_voice_ids.json
. Generated via the Voice Design API and managed with core/voice_design.py
.core/voice_clone_manager.py
) — persistent voiceId
minted from a WAV/MP3 sample.core/voice_cache.py
.VOICE_GENDER_MALE
/VOICE_GENDER_FEMALE
, not "male"
/"female"
strings. Passing the strings 400s silently.Tried on the same sentence, same voice, side-by-side audio comparison:
| Pattern | What it did |
|---|---|
[laugh] [sigh] |
|
| Silence in output | |
(laughs) (sighs) |
|
| Sometimes read literally | |
*laughs* *sighs* |
|
| Silence (asterisks get stripped) | |
<laugh/> <sigh/> |
|
| Silence (not valid SSML on Inworld) | |
<emotion>laugh</emotion> |
|
| Silence |
The Inworld API does not document support for any of these. We had assumed (because every other TTS post on the internet uses them) that they were a universal convention. They are not.
What Inworld does expose is ** temperature** and
speakingRate
After enough A/B-ing across 26 archetypes × 15 languages, four patterns reliably change the audio output.
"You did *what?*"
The asterisks get stripped from the spoken text but the emphasised word lands with audible stress. Works in every voice we tried. The cheapest, highest-hit-rate marker.
"Fine... you can come in."
Three dots produces a real with a tonal drop — the voice equivalent of a sigh, without trying to fake [sigh]
. Five dots for a longer . The model interprets them as prosodic cues.
<break>
for hard s
<speak>
She d. <break time="0.4s"/> "Fine, you can come in."
</speak>
Inworld accepts a useful subset of SSML, and <break>
is the one that matters most for expressive speech. 0.2s
for a beat, 0.4s
for a sigh-, 0.8s
for a beat-before-a-line-delivery moment. Wrap the whole text in <speak>
and the parser handles it.
"Mmm... ha-ha, you're right."
"ahh... I needed that."
The model will render ha-ha
, mmm
, ahh
, oh
, nnn
as the actual sound, because they're spellings of sounds rather than meta-tags. They sound far more natural than a synthesised [laugh]
even when one exists.
For emotional/intimate scenes, rhythmic repeats (ah... ah... ah
) carry actual prosody. We use this for breath patterns where another TTS would want a [breathe]
marker.
In core/voice.py
we run every chunk through enrich_for_tts()
(line ~772) before handing it to Inworld. Regex-based, language-aware, idempotent:
def enrich_for_tts(text: str, lang: str = "en") -> tuple[str, dict]:
"""Return (preprocessed_text, request_params).
Strips fake paralinguistic tags, adds SSML breaks where appropriate,
and bumps temperature/speakingRate for high-emotion scenes."""
text = _STRIP_FAKE_TAGS.sub("", text)
text = _ELLIPSIS_TO_BREAK.sub(r'<break time="0.3s"/>', text)
if "<break" in text:
text = f"<speak>{text}</speak>"
params = _detect_mood_params(text, lang)
return text, params
The mood detector looks for emotional cues (intensity words, repeated punctuation, onomatopoeia density) and bumps temperature
and speakingRate
for the more expressive scenes. Same model, same voice, much more dynamic output, all without any inline tag that the model would have ignored.
[laugh]
/[sigh]
is universal.[sigh]
that emits silence looks identical to one that emits a sigh in any log.temperature
, speakingRate
, and a useful subset of SSML — not inline tags."ahh..."
is a thing the model can read; [sigh]
is a meta-instruction it can't.The audio quality jump from these four patterns is meaningful — users notice. The cost is a 30-line preprocessor and the courage to delete every [laugh]
your team has been sprinkling for months.
This is from production work at ** HoneyChat** — Telegram-native AI companion where voice messages are a first-class output. Canonical version:
— HoneyChat Engineering
temperature
, speakingRate
), SSML subset, voice design API.<break>
, <speak>
, prosody elements.