Lovo AI (Genny) MCP. Generate professional audio from text in a conversation.

Lovo AI (Genny) TTS Server - Voice Synthesis API lets your agent transform plain text into lifelike speech. You'll use this server to generate high-quality, controllable voiceovers for anything—videos, podcasts, or ad copy—by leveraging hundreds of premium voices and granular style controls.

To get started, you first need a handle on what's available. Run list_voices if you wanna see an exhaustive catalog of all the AI speakers we got going. You can filter this massive list by language or gender to narrow down your search immediately.

Once you know which voices exist, you gotta pick the right one. Use get_speaker when you want deep metadata on a specific speaker ID. This pulls detailed information about that voice's capabilities and parameters; it's how you nail the perfect tone for your content before committing to an audio file.

When you're ready, you submit the actual text using create_tts_job. You pass in the source material along with specific synthesis parameters—things like desired speed or emotional style. This job submission doesn't give you the audio; it gives you a unique Job ID. Keep that ID safe because you need it next.

Because generating high-quality speech takes time, you don't get the file right away. You gotta use get_tts_job with your saved Job ID. This function checks the status of the job—it tells you if the audio is still processing in the background, if something failed and needs a retry, or if it’s finally ready for download.

The whole process flows like this: first, run list_voices to browse the field; next, use get_speaker on that ID to fine-tune your requirements; then, feed everything into create_tts_job; and last, monitor the outcome with get_tts_job. You're managing the entire voice production pipeline conversationally through your agent.

For instance, if you need a very specific sound—say, an upbeat corporate announcer in Mandarin—you first run list_voices to find available profiles. Then, you pull up that profile using get_speaker so you can confirm its emotional range and what kind of stylistic controls it supports. Once confirmed, you submit your script via create_tts_job, making sure you set the speed parameter precisely how you want it, along with any required emotional style tags.

The API returns a Job ID immediately. You'll then keep hitting get_tts_job until that status field flips from 'PROCESSING' to 'COMPLETED,' giving your agent the final link to download the finished audio file.