Voicemod launches Key Pocket, bringing voice changing to phones and consoles

Voicemod has launched Key Pocket, a MagSafe device that brings real-time voice changing and soundboards to phones, consoles and social apps without a PC, on sale in the United States with Europe to follow. Article 50 of the AI Act has required providers of systems generating synthetic audio to mark the output in a machine-readable format since 2 August.

Voicemod has launched Key Pocket, a palm-sized device that brings real-time voice changing and soundboards to phones, consoles and social apps without a PC, the company said. It is on sale in the United States, with Europe to follow.

Voicemod is based in Valencia and says 5.7 million people use its software every month. Jaime Bosch, its founder and chief executive, calls voice one of the last untapped layers of digital identity.

The device exists because of an operating system limit. Phones generally do not let one app process microphone audio and feed it into a game or a call in real time, which is why voice changing has stayed on desktops.

Key Pocket works as an audio bridge instead, running alongside the Voicemod app, with processing kept local and separate from the game’s own audio.

It attaches by MagSafe, carries an LED wheel, six programmable trigger buttons and a mute key, takes wired or Bluetooth headsets and passes USB-C charging through.

It also works with PlayStation 4 and 5, Xbox Series X and S, Switch 2 and Meta Quest, reaching the consoles through the controller’s headphone socket.

The catalogue is the draw. More than 200 original voices, 76,000 community-made ones and 800,000 community sounds and memes.

Voicemod’s home market now has a rule about exactly this.

Since 2 August, the AI Act’s Article 50 has required providers of systems that generate synthetic audio to mark the output in a machine-readable format, detectable as artificially generated or manipulated.

It binds providers rather than the people using them, and the carve-out for assistive standard editing does not obviously stretch to changing whose voice you sound like.

Anthropic took the strict reading last month and marks everything, including work the exemption would have let it leave alone.

Voicemod describes its technology as ethical AI. It has not said whether the audio carries a mark, and the announcement does not mention the AI Act at all.

The company raised EUR 14.5M in 2023 and is one of Valencia’s better known exports. Its device is on sale in America, its European launch has no date, and the European obligation is already running.