---
title: "v2026.10.2"
summary: "Dictating no longer wipes the clipboard."
description: "Dictate v2026.10.2 release notes: what changed, and who contributed."
keywords:
  - Dictate
  - v2026.10.2
  - release notes
---

# v2026.10.2

<!-- Written by scripts/release/changelog.mjs from CHANGELOG/2026.10.2.md. Do not edit it: the release cut writes it again. -->

AI agents and tools can read these release notes as [plain Markdown](https://raw.githubusercontent.com/arcforgelabs/dictate/master/CHANGELOG/2026.10.2.md).

This release covers **5 pull requests.**

## Highlights

- Dictating no longer wipes the clipboard. ([#153](https://github.com/arcforgelabs/dictate/pull/153))
- The window connects to the engine in about a second instead of waiting for the speech model to load, and shows "Getting ready…" until it can dictate. ([#158](https://github.com/arcforgelabs/dictate/pull/158))
- Decoding starts the moment the shortcut is released, instead of up to 100 ms later. ([#160](https://github.com/arcforgelabs/dictate/pull/160))
- Each dictation adds one `Dictation timing:` line to the log, with how long each step took and never the text. ([#159](https://github.com/arcforgelabs/dictate/pull/159))

## Changes

- Each dictation adds one `Dictation timing:` line to the log, with how long each step took and never the text. The status line now reads "Transcribing N s of audio", because the old wording looked like a processing time. ([#159](https://github.com/arcforgelabs/dictate/pull/159))

## Fixes

- Dictating no longer wipes the clipboard. Dictate pastes through the clipboard, then puts back exactly what was there before, once the app has read the dictation: text, images, file lists and rich text on Windows, and the copied item on Linux. If you copy something yourself mid-dictation, Dictate leaves it alone. On Windows, dictated text stays out of Win+V clipboard history and cloud clipboard sync. ([#153](https://github.com/arcforgelabs/dictate/pull/153))
- The window connects to the engine in about a second instead of waiting for the speech model to load, and shows "Getting ready…" until it can dictate. A model that fails to load is shown in the window instead of the engine exiting. ([#158](https://github.com/arcforgelabs/dictate/pull/158))
- Decoding starts the moment the shortcut is released, instead of up to 100 ms later. ([#160](https://github.com/arcforgelabs/dictate/pull/160))

## Contribution record

[CHANGELOG/records/2026.10.2.md](https://github.com/arcforgelabs/dictate/blob/master/CHANGELOG/records/2026.10.2.md) lists every pull request in this release once, with its related issues and thanks.