Bolt Voice dictation Turns Spoken Ideas Into App Prompts.

Bolt.new users can now begin building an app by talking rather than typing. The company has launched Voice dictation, a new input method that converts spoken, potentially messy product ideas into a prompt the platform can use to generate an application.

The feature is live now, according to Bolt’s official announcement on LinkedIn. A user selects the microphone, describes what they want to create and lets Bolt interpret the request before beginning the build.

This is not pitched as basic word-for-word transcription. Bolt says Voice dictation removes filler words, repeated points and false starts while accounting for moments when someone backtracks or changes direction halfway through a sentence. The aim is to capture the intended request rather than preserve every spoken word.

Prompt quality remains a practical constraint for AI app builders. A short request may produce a working interface, but it can omit important details about users, workflows, data or design. Writing all of that into a prompt takes time, particularly for founders and small business owners who are still working through the concept.

Voice dictation is designed to reduce that friction. Someone planning an appointment system, for example, could verbally explain who needs access, what information customers submit, how confirmations should work and which administrative controls are required. They can correct themselves during the explanation without manually rewriting the prompt.

Bolt argues that users tend to provide more context when speaking than when filling in a text box. If the system reliably condenses that speech, it could help the first generated version land closer to the original idea. That matters because a clearer starting prompt can reduce the number of follow-up instructions needed to correct basic assumptions.

The company also says users can start speaking in any language. Bolt has not published a detailed list of tested languages, accuracy rates or limitations, so multilingual performance may still vary according to language, accent, microphone quality and the complexity of the request.

The release fits into a broader effort to make Bolt understand more of a project’s requirements before and during generation. The company recently introduced reusable project instructions, as Create With previously reported in its coverage of Bolt.new Skills. Those Skills can carry recurring rules such as preferred technology, fonts and code-review standards across builds.

Voice dictation addresses a different part of the workflow. Skills preserve established preferences, while dictation helps users express an immediate idea with less effort. Used together, they could allow a builder to supply persistent technical constraints in the background and explain a new feature conversationally.

Bolt has also been adding controls around what happens after generation. Its recently launched security agent scans and patches projects before publishing. Dictation moves in the opposite direction, focusing on the earliest stage, when a loosely formed concept becomes an actionable build request.

There are still unanswered operational questions. Bolt’s announcements do not specify whether recordings are retained, how spoken input is processed, whether microphone support differs by browser or whether usage limits apply. No separate price or required subscription tier was disclosed in the supplied materials.

For users, the practical test will be whether Bolt’s cleaned-up interpretation preserves important qualifications while removing verbal clutter. A tool that over-summarizes could quietly discard requirements. One that handles rambling accurately could make prototyping noticeably faster, especially on mobile devices or during early brainstorming.

What is Bolt Voice dictation?

Bolt Voice dictation is an input feature that lets users describe an app idea aloud instead of typing a prompt. Bolt says it removes filler words, repetition and false starts, then builds from the intended meaning of the spoken request.

Which languages does Voice dictation support?

Bolt says users can begin speaking in any language. The company has not provided a language-by-language compatibility list or published accuracy information for different languages and accents.

Is Voice dictation available now, and does it cost extra?

Voice dictation is live in Bolt.new as of August 13, 2026. Bolt has not disclosed a separate price, usage allowance or subscription requirement for the feature in its announcement or blog listing.