Languages

Talk to Kira However You Actually Talk

Most tools quietly force you to think in whatever language the interface was originally built in. Everything downstream from that gets a little clumsier if that is not your first language, or if the people you are actually working with speak something entirely different. A support message gets awkwardly translated before you send it. A call script sounds stilted because it was written for one language and adapted after the fact. An email draft comes out in the wrong register because the model was optimizing for a language that was not the one the recipient would be reading it in.

Kira takes a fundamentally different approach. Language is not a fixed setting you configure once and work around forever. It is something the platform adapts to continuously as your actual conversations happen, matching the real communication in front of you rather than the language your account happened to be set up in.

You set your preferred interface language once inside Settings, and that becomes the language your dashboard, your menus, and your account settings are displayed in day to day. But that is a genuinely separate thing from the language Kira uses when it is working on your behalf out in the world. If you are emailing someone in Spanish, Kira drafts and replies in Spanish, even if your own interface is set to English. If you are on a call with someone speaking French, the call agent adapts to that conversation in real time rather than routing the interaction through a single fixed language regardless of who is actually on the other end.

This distinction matters more than it might initially seem. Teams working across borders, founders managing genuinely international customers, professionals supporting clients who do not share a first language with them, none of these people should have to manually translate everything themselves before Kira can actually help. The language adapts to the real conversation happening in front of you rather than the other way around, where you would be expected to adapt yourself to whatever the software was built to expect.

Why These Languages and Why in This Order

The order in which languages launched is not arbitrary, and it is worth explaining the reasoning rather than leaving it to look like a random prioritization call.

English, Spanish, and French reached full availability first because together they cover an enormous share of Kira's actual early customer base. More importantly, getting the core experience genuinely right in fewer languages first, not just basic translation but real tone-matching, reliable assistant behavior, and consistent product terminology, meant delivering actual quality rather than spreading effort thin across many languages at once and doing all of them adequately instead of a smaller number of them well. A product that handles eight languages at a surface level is less useful than one that handles three with the depth that real work requires.

German, Portuguese, and Italian followed as the next tier, chosen based on where genuine and sustained demand from actual users emerged rather than an internal assumption about which markets seemed strategically important on paper. Real demand from real users is a meaningfully more reliable signal than projected market opportunity, and it guided this decision directly. Hindi and Arabic represent a deliberate and more demanding next step. Both involve fundamentally different script systems and linguistic structures that cannot simply reuse patterns built for European languages. The underlying engineering work is genuinely different in kind, not just more of the same work applied to new vocabulary, which is why these languages take longer to reach the same level of maturity.

Tone Matching Is Not the Same Thing as Translation

This is a distinction that matters enough to explain specifically rather than leaving it implied.

Translation means accurately converting text from one language to another. It is a well understood problem and one that modern AI handles reasonably well across a wide range of language pairs. Tone matching is a meaningfully harder problem. It means studying how a specific person writes, the particular rhythm of their sentences, the words they reach for naturally, the level of formality they default to, and then generating new content that sounds like that specific person rather than like a competent but neutral version of their language.

Email Agent's ability to match your personal writing voice is built on tone matching, not just translation. A language marked Beta in the table below means that basic understanding and response generation work reliably in that language, but the deeper capability of capturing your specific voice and reflecting it back in drafts may still be developing for that language specifically.

In practice, this means an English draft from Email Agent might already sound distinctly like you, capturing your particular cadence and word choices, while a draft in German from the same account might read slightly more neutral until that capability fully matures. This is not a flaw being hidden. It is an honest reflection of how much harder tone matching is than basic translation, and understanding the distinction helps you know what to expect from each language at its current stage of development.

Keeping Kira's Own Vocabulary Consistent Across Languages

Beyond conversational language, there is a quieter but equally important challenge: making sure Kira's own feature names translate and display consistently across every language rather than being rendered slightly differently depending on where they appear.

Smart Calendar, Call Agent, Meeting Homework, Email Agent, these names need to read the same whether you encounter them in the interface, in a notification, in a help article, or in a conversation with the assistant, regardless of which language you are using. We maintain this consistency deliberately, the same way a company maintains a style guide for its product terminology, because without that discipline it is remarkably easy for a product's own vocabulary to drift quietly across different surfaces and languages.

One screen calls it Call Agent. Another translates it slightly differently. A notification uses a third variant. Individually, none of these feels like a significant problem. Collectively, they undermine trust and make the product feel less coherent than it actually is. Consistent terminology across languages is part of the work of building a product that feels genuinely polished in every language it supports, not just the one it was originally designed in.

What Is Supported Right Now

A clarification on what Beta actually means here, since the word gets used loosely in enough places that it has lost precision for many people. Beta does not mean broken. It does not mean unreliable. It means the experience is genuinely functional today and usable for everyday work, while specific aspects of it, particularly things like tone matching described above, are still being actively refined and improved.

German, Portuguese, and Italian interfaces are fully functional right now. The deeper nuance of capturing your specific writing voice in those languages is continuing to mature, but the core experience works reliably.

Hindi and Arabic sit a step earlier in that same process. The assistant can already understand and respond meaningfully in both, currently in beta, while the interface itself is still being built out. This is worth explaining in a bit more detail for Arabic specifically, because the reason it takes longer than other languages is not a simple matter of more translation work.

Arabic is written right to left. That means the entire interface layout needs to mirror correctly, not just translated text supplied into an existing left to right design. Navigation elements, input fields, text alignment, the direction of visual flow across the entire screen, all of it needs to work correctly for an Arabic speaker in a way that goes well beyond what is required for adding a language that shares a left to right script with the languages already supported. That is genuinely more engineering work, and the timeline for Arabic interface support reflects that reality rather than understating the complexity involved.

We made the decision to ship Arabic assistant capability the moment it was genuinely useful rather than holding the entire language back while the interface work continues. A useful assistant in Arabic, even without a fully mirrored interface, is more valuable to someone who needs it today than a complete solution that arrives later. The interface will follow as that work matures.

Vibe Coder's Programming Languages

Language support inside Kira does not stop at how the platform communicates with people. It extends into how Vibe Coder helps you write, edit, and iterate on code.

JavaScript and TypeScript

These cover anything front end, full stack, or built around modern web frameworks. Whether you are sketching out a component, debugging a function, or iterating on a script, these are the languages at the center of most modern web development and they are supported with full depth inside Vibe Coder.

Python

Python covers scripting, automation, data work, and backend development. Python's versatility makes it one of the most commonly reached-for languages across a wide range of contexts, from quick automation scripts to more substantial backend systems, and Vibe Coder's support reflects that range.

HTML, CSS, and JSON

These cover markup, styling, and structured data formats, the foundational layer underneath most of what gets built for the web and the formats that appear throughout modern development workflows even when they are not the primary language being written.

Markdown and Shell Scripts

These cover documentation and quick automation tasks. Markdown is the format that most modern documentation lives in, and shell scripts remain one of the fastest ways to automate repetitive tasks across a development environment.

This list reflects where things stand today. As Vibe Coder matures, expect the supported languages to grow to cover more of what modern development actually touches on a daily basis. The same principle that guides the conversational language rollout applies here: genuine depth in what is already supported takes priority over rushing to add breadth before it is actually ready.

Do Not See Your Language?

If there is a language you genuinely need that is not on this list yet, tell us directly at support@kiraai.ai. Real and specific demand from actual users shapes what gets prioritized next far more meaningfully than any internal assumption made in isolation from what people are actually asking for.

If you are waiting on a specific language for a business reason, a growing customer base in a particular region, a team that operates primarily in a language not yet fully supported, a compliance requirement tied to serving customers in a specific market, mentioning that context when you write in helps us understand not just that a language matters to you but why and how urgently. That specificity makes the signal more useful and more actionable than a general request without context.