Watching an action scene in silence is like eating bread without salt: Music is what gives energy and personality to what we seeBut not everyone knows how to compose a soundtrack from scratch. Google wants to fill that gap, having experimented with musical models within its ecosystem for some time, and is now preparing a more ambitious step with Gemini.
Several recent analyses of the code of the Gemini's Android app is hinting at a new feature capable of generating songs from text.There's no official announcement yet, but the technical details are quite clear: the company is preparing a music tool integrated into the assistant, designed so that anyone can create audio tracks without the need for recording studios, instruments, or complex software.
From isolated experiments to native feature in the Gemini app
Until now, Google had hinted at its interest in AI-generated music in more discreet projects, such as certain features of the Recorder app on the latest Pixel phones or limited tests in combination with video models like Veo. In those cases, AI intervened to accompany images or clips, but there was no musical tool clearly integrated into the user's daily life.
That begins to change with what has been found at Dissecting the APK of version 17.2.51.sa.arm64 of the Gemini app for AndroidThe code contains explicit references to an internal utility named as “MUSIC_GENERATION_AS_TOOL”This name aligns with a new, specific ability to create music within the assistant. The fact that this reference already appears in the application, although not yet visible to the public, indicates that development is in a relatively advanced stage.
Along with this tool, the following have also been detected New references to a category dedicated to music in the "My Stuff" section of GeminiThis area, designed to group what the user generates with AI (texts, images, and other content), would now also include sound compositions. This would make it possible save, organize and play back the generated tracks without leaving the app itself, which is especially convenient for reusing fragments in videos, podcasts, or presentations.
The key novelty is that music would cease to be an isolated experiment and would become a native capabilities beyond those of the assistant, on par with image generation or text writingWithin the Gemini ecosystem, creativity would move across three main fronts: image, written word, and now also audio, all centralized in a single interface.
Lyria's role and Google's previous experience in AI-powered music
This move didn't come out of nowhere. Google has been working with this for some time. specific models for musical composition using artificial intelligence. Among them stands out lyria, a system designed to create instrumental pieces and control their evolution in real time, which until now has been offered mainly through the Gemini API and in pilot projects with developers.
Thanks to Lyria, it is now possible for third-party applications to allow their users generate, direct and play instrumental music interactively, as happened with the AI-created bandHowever, in practice, these capabilities remained somewhat inaccessible to the average user: they required integrating the API, designing a custom interface, and, in many cases, possessing a certain level of technical expertise. The new feature outlined in Gemini would change this approach by bringing this technology directly to the mobile application.
Within the Google ecosystem, Lyria has also been used as a complement to models such as I see, focused on video generationto create soundtracks tailored to AI-generated clips. With musical integration in Gemini, this ability to accompany images with synthetic audio becomes possible. It would be just a tap away.without needing to use separate services or manipulate complex configurations, something that connects with debates about AI problems on YouTube.
The key to this shift lies in the accessibilityInstead of requiring programming or production knowledge, the idea is that anyone can Request songs, melodies, or background sounds using natural language descriptionsFrom content creators looking for background music for their videos to students who want a track to concentrate or do sports, the potential range of uses is wide both in Spain and in the rest of Europe.
From a strategic point of view, Gemini's inclusion of this feature reinforces Google's commitment to an assistant that acts as unified creative centerThe company competes not only in the field of conversational responses, but also in the field of tools that help produce multimedia content, a space where AI-generated music can make a difference compared to other services.
How might generating songs from text work?
Although the final interface has not yet been made public, the references found and the parallels with other Gemini functions allow us to imagine a fairly clear usage scheme. Most likely, the user will have to Write a prompt describing the type of music you wantSimilar to how images are requested. For example: “motivational 80s rock song for working out”, “epic soundtrack for a trailer” or “soft electronic music for studying at night”.
Based on that description, the assistant would combine its ability to interpret natural language with the underlying musical patterngenerating a piece that fits the style, atmosphere, and possibly the specified duration. The experience would be very similar to that of familiar image generation tools: the user types a few lines, waits a few seconds, and receives a ready-to-listen result.
The analyzed code also suggests that there might be basic controls to fine-tune the resultPerhaps adjustments to intensity, predominant instruments, the presence or absence of synthetic vocals, or variations on a single musical idea. Details are not yet confirmed, but it would fit with Google's approach of offering a simple interface with minimal yet practical options.
Another important element is the music category within “My Stuff”. From there, users could save different versions of a song, organize your sound projects, and play them back whenever you wantas if it were a small, personal music library created with AI. This space would also make it easier to rework previous work or share snippets across other services in the Google ecosystem.
Whatever the final design, the intention seems clear: to bring to audio the same immediacy that is already offered in the generation of images and text.If the system proves intuitive enough, many users are likely to integrate music creation into their regular interactions with Gemini, whether to send a message with a personalized melody or to accompany more elaborate projects.
Limitations, access, and possible copyright filters
The text strings found in the application also hint at possible Access restrictions based on account typeIt is possible that the new music tool will initially be linked to paid subscriptions, such as those of Gemini Advanced or Google One plans with expanded AI features, so that free users have limited access or can log in later.
In parallel, everything indicates that they will be implemented Filters to prevent the generation of songs that are too similar to existing commercial works. or the imitation of specific artists' voices without permission. music industry It has been particularly attentive to the emergence of AI for months, and in Europe, moreover, the regulatory framework regarding copyright and data use It requires us to be cautious with these types of tools.
It's reasonable to think that the feature will impose barriers when a user tries to request something like "a song like that artist" or "a voice identical to that singer" without permission; in fact, Google is likely to enforce this. policies against AI-generated music to prevent abuses. EU regulations themselves require transparency in AI models and care with protected content, therefore Google will need to balance creative flexibility with respect for intellectual property. if you want to deploy the tool without legal issues in the European Union.
In addition to the limitations on rights, it remains to be seen whether any will be introduced limits on track duration or on the number of songs that can be generated in a given periodThis is common practice in advanced AI services. These restrictions could serve both to control resource consumption and to manage the load on servers while the feature is being implemented.
For end users, especially content creators in Spain and the rest of Europe, the key point will be understanding Under what conditions can they reuse the music generated by Gemini?If the company clarifies that the tracks can be used freely on social media, podcasts, or commercial projects, the tool could become a very attractive resource compared to traditional audio libraries or third-party solutions.
State of development, deployment and potential impact in Europe and Spain
For the moment, all that is known comes from application analysis and technical leaksGoogle has not publicly announced either the release date or the final appearance of this music feature. The fact that it is already referenced in version 17.2.51.sa.arm64 of the app suggests that the project is active, but also that it is still in the internal testing phase.
It remains to be seen whether the company will opt for a phased deployment by country or whether it will attempt to launch the tool more or less simultaneously in major markets. In the specific case of the European Union, the requirements of the future AI regulatory framework and current intellectual property laws could influence both the timelines and the features available in each region.
For users in Spain, the arrival of this feature would open up an integrated alternative within the Google environment compared to platforms that already offer AI-generated custom tracksThe ability to work directly from the Gemini app, sync creations with a Google account, and have them accessible from different devices would be a practical advantage over scattered external services.
It's still unclear which economic model Google will choose. The company could offer a Free version limited in quality or number of tracksThis could mean reserving the most complete experience for those who pay a subscription, or positioning music generation in the premium features category from the outset. This decision will largely determine whether this technology reaches the general public or remains in the hands of those who already pay for advanced AI tools.
If the integration is ultimately confirmed and the legal issues are resolved, It wouldn't be surprising to see Spanish-language content creators incorporating music generated by Gemini into short videos, podcasts, educational materials, or corporate presentations in the short term.Just as AI assistants are used today to write scripts, summarize documents, or generate images, musical composition could become another step within that digital creative flow.
Everything suggests that this is Google's attempt to make composition accessible to anyone with a mobile phone in their pocket: a function that only requires writing a few lines for Gemini to return a song ready to playprovided the company manages to refine the experience, clarify the legal framework, and decide how to make it available to users in Spain, Europe, and the rest of the world.
