Shareability with piper (Speech models similar to OpenAI but free and self-hosted) #624

New Issue

2025-11-11T14:27:44-06:00

GiteaMirror commented

2025-11-11 14:27:44 -06:00

Originally created by @HirCoir on GitHub (Apr 10, 2024).

Please add shareability with piper tts, I leave you a repo with different branches Main (Linux Aoo Flask), api-server and Flask App for windows:
https://tts.hircoir.eu.org/
https://github.com/HirCoir/HirCoir-Piper-tts-app/tree/piper-api-flask
It would work this way:

const audioResponse = await axios.post(¨https://piper-low.studio.hircoir.eu.org/¨, {
          text: assistantResponse,
          model: ¨kamora¨
        }, {
          headers: {
            'Content-Type': 'application/json'
          }
        });

        const audioBase64 = audioResponse.data.audio_base64;
        const audioBuffer = Buffer.from(audioBase64, 'base64');

        await sendAudio(client, message, audioBuffer);
      } catch (error) {
        console.error('Error al realizar la solicitud:', error);
        await client.sendText(message.from, 'Error.');
      }
    }

Piper has countless models trained by the community and are available on Huggingface: https://huggingface.co/models?sort=trending&search=piper
The quality is similar to an OpenAI model. Even the user can create his own model by following these steps:https://ssamjh.nz/create-custom-piper-tts-voice/

Originally created by @HirCoir on GitHub (Apr 10, 2024). Please add shareability with piper tts, I leave you a repo with different branches Main (Linux Aoo Flask), api-server and Flask App for windows: https://tts.hircoir.eu.org/ https://github.com/HirCoir/HirCoir-Piper-tts-app/tree/piper-api-flask It would work this way: ``` const audioResponse = await axios.post(¨https://piper-low.studio.hircoir.eu.org/¨, { text: assistantResponse, model: ¨kamora¨ }, { headers: { 'Content-Type': 'application/json' } }); const audioBase64 = audioResponse.data.audio_base64; const audioBuffer = Buffer.from(audioBase64, 'base64'); await sendAudio(client, message, audioBuffer); } catch (error) { console.error('Error al realizar la solicitud:', error); await client.sendText(message.from, 'Error.'); } } ``` Piper has countless models trained by the community and are available on Huggingface: https://huggingface.co/models?sort=trending&search=piper The quality is similar to an OpenAI model. Even the user can create his own model by following these steps:https://ssamjh.nz/create-custom-piper-tts-voice/

GiteaMirror closed this issue

2025-11-11 14:27:44 -06:00

GiteaMirror commented

2025-11-11 14:27:45 -06:00