Changelog

v0.9.1

19 October 2026, Latest release

Kimiko in eight languages

New

• Kimiko in your language. Menus, buttons and messages now come in English, German, French, Spanish, Italian, Polish, Ukrainian and Russian. Kimiko follows your system language, or pick one in Settings → General → App language.

Improved

• A tidier providers list. In Settings → AI, each provider takes one line with its model and key status, and one Add provider menu lists them all.
• On macOS 27, the accessibility permission uses the system’s name, Device Control and Data Access.

v0.8.3

16 October 2026

Summaries in your meeting’s language

Improved

• Summaries follow the spoken language. With Same as meeting, summaries, titles and proposed tasks come out in the language the meeting was held in, including with the built-in models.
• A language for each meeting. Pick one in the enhanced notes menu or beside live summaries, and Kimiko keeps it for that meeting.
• The chat box stays at the bottom when you start a new conversation.

After this update, earlier versions of Kimiko can’t open your notes. Nothing is deleted.

Fixed

• Built-in models now name task owners, and Gemma 4 E2B adds due dates.
• Summaries are less likely to credit you with what others said.
• Renaming a speaker updates names on indented bullet points too.

v0.8.2

15 October 2026

A fresh start in Chat and three new AI providers

New

• A fresh start in Chat. New conversations open with a greeting and two suggested questions to start from.
• Stop an answer at any time with the Stop button or Esc, in Chat and in a note’s chat.
• Mistral, Muse by Meta and Grok. Add them with your own API key in Settings → AI → Custom providers. Mistral and Meta models that may train on your content are left out.

Improved

• Keep typing while Kimiko answers. Your next question waits and sends when the answer is done.
• Answers keep going when you switch to another note or page.
• The model button shows which model Auto uses and warns when it can’t run.
• Conversations are grouped by date, and @ opens a compact picker for notes and projects.

Fixed

• Chat references show the right speaker names.
• A question that fails goes back into the chat box so you can resend it.
• Mistral’s reasoning models no longer stop with an error.

Ask questions about your notes

v0.8.0

13 October 2026

A new search model for chat

Improved

• Better chat search. Chat now searches with Google’s EmbeddingGemma 2 on your computer and reads meeting titles too, so asking about a meeting by name works better.

After the update, Kimiko downloads the new model (about 310 MB) and re-indexes your meetings in the background. Until then, chat uses keyword search.

Ask questions about your notes

v0.7.21

13 October 2026

A new Voxtral model and a quicker start on Mac

Improved

• Voxtral Mini 4B Realtime replaces Voxtral Mini 3B. It’s built for transcription and covers 13 languages, adding Russian, Chinese, Japanese, Arabic and Korean.
• A quicker first start on Mac. The first recording after an update starts in about 5 seconds instead of more than 20.

Kimiko removes Voxtral Mini 3B, freeing nearly 3 GB. To keep using Voxtral, download the new model in Settings → Transcription → Speech models.

Fixed

• Transcripts keep the speech before a model gets stuck repeating a phrase.

Compare speech models

v0.7.20

11 October 2026

A downloaded update stays ready

Fixed

• Restart to update stays available after a failed update check.
• A waiting update is no longer downloaded again.

v0.7.19

11 October 2026

Tasks in your meeting’s language

Improved

• Proposed tasks in your meeting’s language, like summaries.
• One language setting. Settings → AI → Summary & task language now covers both summaries and tasks.

Fixed

• Open in transcript appears only on tasks that have a moment to jump to.

Keep track of tasks

v0.7.18

10 October 2026

Simpler, clearer Settings

Improved

• Simpler Settings. Shorter pages and nine sidebar items in four groups. Connections holds your calendar and AI apps, and Privacy & data holds Backup and Trash.
• One button per AI app to connect Claude, Claude Code, Cursor or VS Code.
• See which AI model is in use, and why summaries can’t run when something is missing.

Fixed

• Switching between Built-in AI and your own providers keeps your model choices.
• Privacy notices now say exactly what goes to cloud providers.

v0.7.17

8 October 2026

Claude Haiku 5.5 on OpenCode Go

Fixed

• Claude Haiku 5.5 and more OpenCode Go models, such as Grok, MiniMax and GPT-5.6 Luna, now work.
• Summarize into notes no longer cuts summaries short with models that think first.

Check your cloud provider setup

v0.7.16

8 October 2026

Quicker retries for update checks

Improved

• A failed update check tries again after 2 and 10 minutes instead of six hours.

Fixed

• Deactivate tells you when this computer wasn’t released from your license.

v0.7.15

7 October 2026

A shorter setup and a Permissions page

New

• A shorter setup in six quick screens. The speech model downloads while you finish, and built-in AI is optional.
• Nothing starts before you do. Kimiko downloads, records and checks for updates only after you choose Get started.
• Settings → Permissions shows what Kimiko can access and helps you fix anything that’s off.
• Connect a calendar from Coming up, with a clear notice of what Google sign-in reads.

Improved

• Calendar settings show when your calendar last synced, and sign-in problems get plain messages.

Allow microphone and system audio

v0.7.14

5 October 2026

Fewer license interruptions

Improved

• Kimiko checks your license about once an hour. Use the refresh button in Settings → License to check right away.

Fixed

• A busy license service no longer stops recording or rejects a valid key.

Use Kimiko offline

v0.7.13

3 October 2026

Speaker labels for imported recordings

New

• Who said what in imports. Imported files and links get speaker labels for up to eight voices. Turn it off with Label speakers.
• Re-transcribe and Re-diarize imports when Keep session audio is on.

Improved

• Re-diarize uses the same speaker model as live meetings, on Mac and Windows.

v0.7.12

2 October 2026

A tidier License page

Improved

• Settings → License shows one short line per row, with the details behind ⓘ.

Fixed

• The trial countdown no longer shows “0 days” when only meetings remain.

v0.7.11

1 October 2026

Manage your license on kimiko.app

New

• Subscribe and manage billing on kimiko.app, straight from Settings → License.
• Manage computers and find your key in your Kimiko account.
• Easier activation, with a quick way to free a spot when your key is on 3 computers.

Improved

• License changes reach Kimiko faster, with clearer wording.

Fixed

• The recording-paused screen gives the real reason and offers Re-activate when that’s what you need.

Manage your license and computers

v0.7.10

27 September 2026

Tidier task proposals

Improved

• Proposed tasks take one line in the Tasks pane. Open one to read it in full.

Fixed

• Task windows show long titles and empty details correctly.

v0.7.8

26 September 2026

An updated speech engine

Improved

• An updated on-device engine for the Parakeet, Nemotron 3.5 Streaming, Qwen3 ASR, Voxtral and Cohere Transcribe speech models.

On a Mac, the first recording after the update takes longer to start.

v0.7.7

26 September 2026

Graphics memory freed after meetings on Windows

Fixed

• On Windows, Kimiko frees graphics memory as soon as you stop recording.

v0.7.5

25 September 2026

Speaker labels for larger meetings

Improved

• Up to eight remote speakers, labeled on your device by NVIDIA’s Nemotron 3 Diarization (about a 401 MB download).
• Long titles wrap instead of being cut off.

Fixed

• Activating a license ends the trial on that computer.

v0.7.4

24 September 2026

A tidier toolbar in smaller windows

Fixed

• In narrow windows, the formatting toolbar folds into a menu instead of covering the side pane.

v0.7.3

22 September 2026

Give the conversation more room

New

• Expand the meeting pane to give the transcript, live summaries or tasks the whole window (⌘⇧M on Mac, Ctrl+Shift+M on Windows).

Fixed

• Pinned chat keeps unsaved edits, and the sidebar no longer jumps.

v0.7.2

21 September 2026

More reliable audio imports

Improved

• Security updates for cloud connections and stronger local-only safeguards.

Fixed

• MP3 imports keep their exact length and timestamps.

v0.7.1

20 September 2026

Find the details in your transcripts

Fixed

• Chat now finds details in transcripts, not only in summaries. Older meetings are repaired in the background.

Ask questions about your notes

v0.7.0

19 September 2026

More control over imported recordings

New

• Choose the language for a link import.
• Import straight into a project from its page.

Improved

• Cancel a link import at any point before the note is saved.

Fixed

• Imports follow your transcription language and the recording’s original audio track.

v0.6.8

19 September 2026

Turn a recording link into notes

New

• Import from link. Paste a recording link on the Notes or a project page to get a transcript and summary, with a link back to the source.
• MP3 imports.

Link imports need an internet connection. Recordings behind a login, live streams, playlists and DRM-protected media aren’t supported. Only import recordings you’re allowed to use.

Improved

• Long audio files use less memory while importing.

v0.6.7

16 September 2026

More control over dictation memory

New

• Keep models loaded. Choose in Settings → Dictation how long dictation models stay ready between uses.

Improved

• Dictation models unload after five minutes by default, and AI formatting starts loading as you dictate.

Adjust dictation memory

v0.6.6

16 September 2026

More room for meetings on laptops

Improved

• Kimiko frees dictation memory before a meeting starts and shares graphics memory better on Windows laptops.
• A simpler indicator shows whether transcription runs on the CPU or GPU.

Understand memory use on laptops

v0.6.5

15 September 2026

Less graphics work on Windows

Improved

• On Windows, speech detection moves to the main processor, and so does speaker labeling on computers with shared-memory graphics.
• Floating chat no longer dims the document behind it.

Choose a setup for your computer

v0.6.4

13 September 2026

Lower resource use after meetings

Improved

• Background AI models free memory soon after they finish. Models you’re chatting with stay ready.

Fixed

• On Windows, the graphics processor no longer stays busy after a meeting.

Read about local AI

v0.6.2

10 September 2026

Clearer help when AI tasks fail

Improved

• Failed summaries show what went wrong and a Retry button, even after a restart.

Fixed

• More reliable OpenCode Zen and Go connections.

Check your cloud provider setup

v0.6.1

9 September 2026

Speaker names stay in sync

Improved

• Rename a speaker in the summary or transcript, and every linked name in that meeting follows.

Work with speaker labels

v0.6.0

9 September 2026

Project notes and tasks, together

New

• Projects show their notes and tasks together, for the whole project or one note.
• Notes and recordings you start from a project are filed into it.

Improved

• Faster, more resilient model downloads, and easier list editing in notes.

Keep track of tasks

v0.5.9

7 September 2026

More room for the conversation

Improved

• The meeting pane and task rows make better use of smaller windows.
• License reminders stay quiet while you record or dictate.
• The Windows installer uses Kimiko’s official artwork.

Explore meeting notes

v0.5.8

6 September 2026

Smoother dictation on Windows

Improved

• A new Windows shortcut. Dictation starts with Ctrl + Win by default.
• Dictation starts sooner, with an optional sound when the microphone is ready.

Fixed

• Dictated text lands where you’re typing on Windows, including Notepad and Kimiko’s own notes.

Set up voice dictation

Get the latest version

Open Kimiko’s update controls in Settings, or download the latest Mac or Windows installer.