On January 28, 2026, reports emerged that Google is internally testing a significant update for Google AI Studio that would allow developers and users to clone their own voices.
The feature, tentatively discovered as a hidden “Create Your Voice” option, signals Google’s move toward native, high-fidelity audio personalization within its Gemini ecosystem.
1. The “Create Your Voice” Discovery
The feature was first spotted by researchers at TestingCatalog and confirmed by tech outlets on January 28.
- The Trigger: The option briefly appears in the Google AI Studio interface when a user selects the “Flash Native Audio Preview” model (currently associated with Gemini 2.5 Flash).
- The Interface: Clicking the option opens a dedicated pop-up window that prompts the user to either record live audio or upload an existing audio file of their voice.
- Current Status: As of late January 2026, the button is inactive for most users, suggesting Google is in the “UI-testing” phase before a full backend rollout.
2. Powered by Gemini 3 Flash?
Industry analysts believe this feature is the precursor to the launch of Gemini 3 Flash, which is expected to feature “Native Audio” as a core capability.
- Native Audio vs. TTS: Unlike traditional Text-to-Speech (TTS) that layers a voice over a model’s output, a Native Audio model processes and generates sound directly, allowing for much higher emotional nuance and lower latency (~97ms).
- Instruction Following: The new model is expected to allow users to not only clone a voice but also give natural language instructions like “Say this in my voice, but sound excited and speak faster.”
3. Key Potential Use Cases
If released, direct voice cloning in AI Studio would provide a massive boost to several developer-led sectors:
- Accessibility: Creating personalized digital voices for individuals with speech impairments.
- Gaming & Apps: Developers can build NPCs (Non-Player Characters) or virtual assistants that use the user’s own voice (with consent).
- Content Creation: Podcasters and YouTubers could “fix” audio errors by typing corrections that are then spoken by their cloned AI voice.
4. Other Upcoming AI Studio Features
The voice cloning discovery was part of a larger set of unannounced “Power User” tools found in the January 2026 backend update:
- Import from GitHub: A new button in the “Build” section will reportedly allow developers to pull entire repositories into AI Studio as active projects.
- Separated Usage Stats: The homepage is being redesigned to separate “Activity Summaries” from “Usage Statistics” for better project management.
- Firebase Integration: Deeper backend links to Firebase are expected to launch alongside the voice tools to help “full-stack” AI apps go live faster.
Conclusion: The End of Third-Party Voice Middleware?
By integrating voice cloning directly into AI Studio, Google is positioning itself to compete with specialized platforms like ElevenLabs and Veed. If the “Gemini 3 Flash” rollout includes this feature for free (within API limits), it could drastically lower the cost and technical barrier for building audio-first AI applications in 2026.
