Google gives the Gemini app an ear for audio files

For months, the biggest missing piece in Google’s flagship chatbot wasn't a smarter reasoning engine or a slicker interface—it was the simple ability to drop in a voice memo.

That finally changed on Monday. Josh Woodward, vice president of Google Labs and Gemini, posted on X that audio upload support has officially arrived on the Gemini app, calling it the platform's "#1 request."

Drop in the audio, but watch the clock

The feature brings direct voice and recording analysis into Gemini prompts, though Google is keeping free users on a tight leash. If you don't pay for Gemini, you can only upload clips up to 10 minutes long, capped at five free prompts per day.

Power users willing to pay for AI Pro or AI Ultra plans get a much wider runway, with support for single audio files up to three hours long. All tiers can upload up to 10 files per prompt across various formats, and Gemini can even unpack files straight out of uploaded ZIP archives.

NotebookLM makes flashcards, Search adds tongues

The audio rollout is part of a broader triple-threat update across Google's AI lineup. NotebookLM, which already supported audio as a research tool, now lets users turn their uploaded notes and files into custom reports across more than 80 languages.

Instead of just generic summaries, NotebookLM can now output formatted study guides, executive briefings, blog posts, flashcards, and quizzes, with full control over style and tone. Google noted on X that this feature should be 100% rolled out by the end of this week.

Meanwhile, Google Search's AI Mode integrated Gemini 2.5 to expand into five new languages: Hindi, Indonesian, Japanese, Korean, and Brazilian Portuguese. It caps off a breathless stretch for Google, following August updates that added persistent conversation memory to Gemini and free access to Workspace Vids, along with September's Photos upgrade to the Veo 3 video model.

Why it matters

If you regularly deal with recorded lectures, podcast transcripts, or customer calls, this instantly eliminates an annoying middle step. You no longer have to transcribe audio through a third-party app just to paste raw text into your chatbot.

That said, Google’s tiering here is brutally calculated. A 10-minute cap with five daily turns makes the free tier little more than a demo for casual voice notes. If you want to process actual workplace meetings or full-length interviews, you have no choice but to pony up for a paid tier. We think it is a smart commercial play to monetize Gemini's multimodal advantage, though users accustomed to generous free allowances might find the daily prompt limits restrictive.

Across the broader landscape, Google is finally connecting the dots of its ecosystem. Rather than trapping multimodal magic inside niche research experiments like NotebookLM, bringing native audio directly into Gemini and Search turns Google's multimodal muscle into an everyday enterprise-ready utility.