Add a voice to your bot

Click a pack or a voice to preview it. The download is one compressed package: the voice records, one synthesizer, and a short README. You pay once. There is no monthly fee. Open it on CPU. No GPU. One loaded voice uses about 140 MB of RAM, and about 180 MB while a short line is speaking.

Put it in your app

  1. Download the pack. The .ahv9 inside is the compressed synthesizer.
  2. Open that file on CPU with AhanaTool. No account and no GPU.
  3. Call speech from your app: POST /v1/audio/speech with {"input":"Hello","voice":"PACK-small","engine":"own"}.

One voice, with its own synthesizer, is $7.77. A pack of 5 or more is $5.00 a voice into one custom pack that shares a single synthesizer. The custom pack starts at 5 voices.

Pack size

Measured files, ready to send. A one-voice package is 58,920,960 bytes. Shared packs stay near 59 MB because they keep one synthesizer.

VoicesShared pack fileEach voice its own fileSharing saves
559,002,880294,604,800235,601,920
1059,013,120589,209,600530,196,480
2559,023,3601,473,024,0001,414,000,640
5059,054,0802,946,048,0002,886,993,920

Shared synthesizer

$5.00 a voice · 5 or more

Opens on CPU. Five-voice file: 59,002,880 bytes. Fifty-voice file: 59,054,080 bytes.

CPU: 1 core. About 12 s to open. About 0.1 s for a short line after that.

GPU: not used.

RAM: about 140 MB loaded. About 180 MB speaking. Same for 5 through 50 voices.

Own synthesizer

$7.77 a voice

Opens on CPU. One file: 58,920,960 bytes. Five files: 294,604,800 bytes. Fifty files: 2,946,048,000 bytes.

CPU: 1 core for each synthesizer that stays loaded.

GPU: not used.

RAM: about 140 MB each. About 180 MB while that voice speaks. Five together is about 700 MB.

Packs

Click the card to hear the packaged voice. Download sends the premade tar.

Country packs

Five voices from one country, one synthesizer, $25 once.

À la carte

Pick at least 5 voices. $5 each. They ship as one package with one synthesizer. Click a card to preview that voice.

0 selected