v2026.10.11
Release date: October 6, 2026
New Features
Gemini for conversation and voices
Google Gemini is available for conversation and text-to-speech. The same key serves the agent and voices. Without a key, catalog models use the AI credit balance.
What it is for
- Answer with Gemini models on the agent, prompt test, and copilot
- Generate audio with 30 built-in voices and five TTS models
- Hear a sample with the chosen text and model before using the voice in chat
How to use it
- Under Settings → Integrations, save a Google AI Studio key, or use Interflow credits
- On the agent, pick the Gemini integration and the model. Under Parameters, set the reasoning level (Low, Medium, or High) and the max tokens. Temperature does not apply to these models
- Under Voices → Catalog, open Gemini or OpenAI, click Test text and, if you want, change the model inside the test. Use this voice saves it on the Interflow balance. The OpenAI default is GPT-4o Mini TTS
- Edit the suggested text and generate. The file is MP3 and spends credits when the voice uses the Interflow balance
- Use this voice saves the preset under My voices. The same test is on the card and in the form, including ElevenLabs
Full guide: Gemini integration · System voices
Benefits
- Gemini conversation and voice on the same key, or on the credit balance
- A sample with the chosen model and text before sending audio to the customer
- The file arrives as MP3, ready for the conversation
- In the copilot, the Gemini reply streams in, and the reasoning summary follows the low, medium, or high level