Rumik AI

Text-to-speech service implementations using Rumik AI’s TTS APIs

Overview

pipecat-rumik provides Pipecat text-to-speech services backed by Rumik AI’s TTS APIs. It exposes RumikTTSService for WebSocket-based synthesis in interactive voice pipelines and RumikHttpTTSService for HTTP request/response synthesis. Both services emit raw PCM audio frames (TTSAudioRawFrame).

Source Repository

Source code, examples, and issues for the Rumik AI integration

PyPI Package

The pipecat-rumik package on PyPI

Rumik AI

Learn more about Rumik AI

API Keys

Create and manage your Rumik AI API keys

Installation

This is a community-maintained package distributed separately from pipecat-ai:

pip install pipecat-rumik

Prerequisites

Rumik AI Account Setup

Before using either service, you need:

1. API Key: Create API keys from the Rumik AI dashboard
2. Gateway URL: Use the gateway URL provided for your Rumik AI deployment

Required Environment Variables

• RUMIK_API_KEY: Your Rumik AI API key
• RUMIK_GATEWAY_URL: Your Rumik AI gateway base URL

Configuration

RumikTTSService

api_key · str · required

Rumik AI API key.

gateway_url · str · required

Rumik AI gateway base URL.

settings · RumikTTSService.Settings · default: "None"

Runtime-configurable TTS settings. See Settings below.

sample_rate · int · default: "24000"

Output audio sample rate. Rumik currently returns 24 kHz PCM.

full_response_aggregation · bool · default: "True"

When true, buffer a complete LLM response before sending text to Rumik instead of creating a separate TTS request for every sentence.

RumikHttpTTSService

api_key · str · required

Rumik AI API key.

gateway_url · str · required

Rumik AI gateway base URL.

aiohttp_session · aiohttp.ClientSession · required

Caller-owned HTTP session used for Rumik API requests.

settings · RumikHttpTTSService.Settings · default: "None"

Runtime-configurable TTS settings. See Settings below.

sample_rate · int · default: "24000"

Output audio sample rate. Rumik currently returns 24 kHz PCM.

Settings

Both services use Pipecat’s service settings pattern. RumikTTSService.Settings and RumikHttpTTSService.Settings are aliases of RumikTTSSettings.

Parameter	Type	Default	Description
model	str | None	"muga"	Rumik model identifier.
voice	str | None	None	Preset speaker voice. Sent to Rumik as speaker.
language	Language | str | None	None	Reserved for Pipecat provider compatibility.
description	str | None	None	Natural-language voice/style description for expressive models.
f0_up_key	int | None	None	Pitch shift in semitones for preset speaker voices.
temperature	float | None	None	Sampling temperature. When omitted, Rumik uses its API default.
top_p	float | None	None	Nucleus sampling value. When omitted, Rumik uses its API default.
top_k	int | None	None	Top-k sampling value. When omitted, Rumik uses its API default.
repetition_penalty	float | None	None	Penalty applied to repeated tokens. When omitted, Rumik uses default.
max_new_tokens	int | None	None	Maximum generated audio tokens. When omitted, Rumik uses its default.

See the source repository for the authoritative, up-to-date list of settings.

Usage

WebSocket Service

import os

from pipecat_rumik import RumikTTSService

tts = RumikTTSService(
    api_key=os.environ["RUMIK_API_KEY"],
    gateway_url=os.environ["RUMIK_GATEWAY_URL"],
    settings=RumikTTSService.Settings(
        model="muga",
    ),
)

For expressive models, use the voice setting for preset speaker voices. The service sends it to Rumik as speaker.

tts = RumikTTSService(
    api_key=os.environ["RUMIK_API_KEY"],
    gateway_url=os.environ["RUMIK_GATEWAY_URL"],
    settings=RumikTTSService.Settings(
        model="mulberry",
        voice="speaker_1",
        description=(
            "warm expressive Indian woman with clear Hinglish diction, natural "
            "pauses, gentle energy, and a friendly conversational delivery"
        ),
        f0_up_key=3,
    ),
)

HTTP Service

import asyncio
import os

import aiohttp

from pipecat_rumik import RumikHttpTTSService


async def main():
    async with aiohttp.ClientSession() as session:
        tts = RumikHttpTTSService(
            api_key=os.environ["RUMIK_API_KEY"],
            gateway_url=os.environ["RUMIK_GATEWAY_URL"],
            aiohttp_session=session,
            settings=RumikHttpTTSService.Settings(
                model="mulberry",
                voice="speaker_2",
                description="calm female narrator",
            ),
        )


asyncio.run(main())

Compatibility

Tested with Pipecat v1.3.0 and supports Pipecat >=1.0.0,<2. Check the source repository for the latest tested version and changelog.