How to Dictate in Linear & Jira by Voice (2026)

TL;DR: Neither Linear nor Jira has native desktop dictation for writing issues. Their AI features generate and summarize text; they do not transcribe your voice. To draft tickets by voice, run a system-wide Mac dictation tool like Voibe: put your cursor in any field — title, description, acceptance criteria, a comment, a sub-task — hold your hotkey, and speak. Because it runs on-device, ticket content stays on your machine, and a custom vocabulary makes your product, component, and epic names transcribe correctly every time.

If you write a lot of tickets, this is the highest-leverage place to use voice on a Mac: an issue is mostly prose — a title, a description, and structured acceptance criteria — and prose is exactly what dictation is faster at than typing. This guide covers where you can dictate in both tools, the native reality, the system-wide setup, custom vocabulary for internal names, a repeatable Voice Ticket Template, and real examples for a bug report, a feature ticket, and a standup comment.

Both tools are mostly text fields, and a system-wide dictation tool works in all of them because it inserts text wherever the cursor is — in the desktop apps and in the browser. Here is the coverage across both:

Neither tool restricts where a system-wide tool can type, because to macOS these are ordinary text inputs. That means one hotkey covers the Linear desktop app, the Jira web editor, and — critically — everything outside them: the Slack thread where the bug was reported, the PRD in your docs tool, the email to a stakeholder. You dictate the same way everywhere.

It is worth being precise here, because both tools ship prominent AI features that sound adjacent to dictation but are not.

Linear has an AI layer — Triage Intelligence, Product Intelligence, and issue generation. On issue creation it can expand a short title into a detailed description with suggested acceptance criteria, and it can read project updates back to you as an audio digest. All of that is text generation and audio playback; none of it is speech-to-text for writing an issue. You cannot speak a ticket into Linear.

Jira has Atlassian Intelligence and Rovo, which generate, summarize, and rewrite the content of work items and comments. Jira historically exposed the browser's speech-recognition button in its editor, but that is no longer a standard feature in current Jira Cloud. Rovo Desktop can use your microphone for dictation, but only to write prompts to the Rovo AI — not to fill in an issue's title, description, or acceptance-criteria fields.

So the gap is real and symmetric: to get spoken words into an actual issue field in either tool, you need a system-wide dictation tool sitting above the app. The upside is that one tool then covers both trackers identically — and every other app you touch while writing tickets.

Any system-wide Mac dictation tool will type into Linear and Jira. This guide uses Voibe because it runs on-device — which matters for ticket content that names unreleased features — and because its custom vocabulary handles the internal product and component names that fill your tickets.

1. Download Voibe from getvoibe.com (or the direct .dmg) and drag it to Applications.
2. Launch it. On Apple Silicon (M1–M4, macOS 13+) it downloads a local Whisper model (~2 GB) on first run.
3. No account, and no internet needed after the model downloads.

For the full install walkthrough, see our Voibe setup guide.

A system-wide tool needs two macOS permissions to type into Linear and Jira:

1. Accessibility — to insert text into the fields (System Settings > Privacy & Security > Accessibility, then enable the app).
2. Microphone — to capture speech (granted on first use, or under System Settings > Privacy & Security > Microphone).

Pick a hotkey you can hold while your hands rest on the keyboard. Voibe defaults to holding Fn: press and hold, speak, release, and the text appears at the cursor. Reassign it in settings if Fn clashes with your layout — see our Mac dictation keyboard shortcuts guide for options and conflicts. Hold-to-talk suits ticket writing because you keep one hand on the key, speak a description, release, and your hands are already back on the keyboard to tab to the next field.

This is the step that makes voice ticket-writing actually reliable, and it is the one thing a general speech model can't do on its own. Every ticket is full of names the model has never seen: your product line, your services, your components, your epics, your team's jargon. Say them, and plain dictation guesses — it splits CheckoutService into "checkout service," writes "Rovo" as "rover," and turns your "Aurora epic" into "a roar a epic."

Open Voibe's settings and add these terms to your custom vocabulary once: product names, component and service names, epic and initiative names, key acronyms, and any teammate names you @-mention often. From then on they transcribe correctly every time you dictate a title, description, or comment. Voibe's custom vocabulary influences the transcription itself rather than doing a blunt find-and-replace afterward, so a service like PaymentsGateway lands as the real identifier, not a phonetic guess.

Engineers writing tickets that reference real code can go further: Voibe's Developer Mode detects an open editor and resolves workspace and project terms — file, folder, and service names — so a bug ticket that mentions authMiddleware.ts gets the exact spelling from your project rather than the literal words.

Dictation rewards structure. If you speak a ticket as one long run-on, transcription drops words and the ticket reads like a stream of thought. Instead, dictate each part of a ticket as its own short burst, in a fixed order. Call it the Voice Ticket Template — four parts you speak in sequence, tabbing between fields:

1. Title — one line, action-first. "Checkout page hangs when the cart is empty."
2. Context / user story — the why, in As a / I want / so that form. "As a shopper, I want the checkout page to load with an empty cart, so that I can add items without a dead end."
3. Acceptance criteria — the done definition, in Given / When / Then form, one criterion per burst. "Given an empty cart, when I open the checkout page, then I see an empty-state message with a link back to the catalog."
4. Notes — repro steps, links, component, or priority. "Reproduces on the CheckoutService in the Aurora epic; blocks the release."

Speaking the same four parts in the same order every time does two things: it keeps each dictation burst short enough to transcribe cleanly, and it produces consistent, well-formed tickets your team can read at a glance. The As a / I want / so that and Given / When / Then patterns are especially voice-friendly because they are formulaic — you fill in the blanks by speaking.

Here is what the Voice Ticket Template sounds like in practice across Linear and Jira. Hold your hotkey, speak each part, release, tab to the next field.

A Bug Report

In the title field: "Empty cart crashes the checkout page." Then in the description:

• "As a shopper, I want to open the checkout page with an empty cart, so that I can start adding items without hitting an error."
• "Given an empty cart, when I navigate to the checkout page, then the page loads an empty-state message instead of a blank screen."
• "Reproduces every time on the CheckoutService in the Aurora epic. Priority high, blocks the release."

A Feature Ticket with Acceptance Criteria

Title: "Add server-side pagination to the users table." Description:

• "As an admin, I want the users table paginated, so that large teams load quickly."
• "Given more than twenty users, when I open the users table, then I see twenty per page with next and previous controls."
• "Given I am on page two, when I refresh, then I stay on page two."

A Standup or Status Comment

In a comment or status update: "Update: the PaymentsGateway migration is code-complete and in review. Testing tomorrow, on track for the Aurora release."

Keep each burst short and structured — voice handles a focused one-to-two-sentence chunk far better than a long run-on. For a repeatable structure to dictate prompts by when you hand a ticket to an AI assistant, use the Five-Part Voice Prompt framework in our voice-prompt AI guide.

• Load your custom vocabulary first. Product, component, service, and epic names are what dictation gets wrong. Add them once and stop fixing them per ticket.
• Dictate one part of the template at a time. Title, then user story, then each acceptance criterion as a separate burst — short chunks transcribe most accurately.
• Say the formula out loud. "As a, I want, so that" and "Given, when, then" are formulaic; speaking the connective words keeps the structure intact.
• Let Developer Mode handle real code names. If a ticket references files or services in an open editor, workspace resolution spells them exactly.
• Speak state, then action, then outcome. For acceptance criteria, that order matches Given/When/Then and reads cleanly.
• Use a decent microphone. A basic USB mic noticeably reduces errors on technical terms and acronyms.
• Edit with the keyboard. Voice is fastest for the first draft of a description; fix the last few words by hand before you submit.

Text isn't appearing in the field

This is almost always a missing Accessibility permission. Open System Settings > Privacy & Security > Accessibility and confirm your dictation app is enabled. If it is on but still not typing, remove and re-add it to reset the permission — this fixes most cases after an app update.

It types into the wrong field

Click directly into the target field first so the cursor is there before you hold your hotkey. In Jira's rich-text editor and Linear's description, make sure the field has focus rather than the surrounding page.

Component and product names are mis-transcribed

Add them to your custom vocabulary. General speech models don't know "PaymentsGateway," "Rovo," or your epic names until you register them. This is the single biggest accuracy win for ticket writing.

Real code names aren't resolving

Confirm Developer Mode is on and your editor is open with the project loaded. Resolution matches terms in the open workspace; add the term to your custom vocabulary as a fallback.

The web app behaves differently from the desktop app

A system-wide tool inserts text the same way in both, but some editors intercept certain keys. If a field misbehaves in the browser, try the desktop app (or vice versa) — the same hotkey and vocabulary carry over.

• Voibe — system-wide and on-device, with custom vocabulary for your product and component names and Developer Mode for real code terms. Works in Linear, Jira, the web app, and every other app. $149 lifetime or $7.50/month, 7-day trial, no account. Best fit for high-volume ticket writing on Mac. See our best dictation software for developers guide.
• Apple Dictation — free, system-wide baseline, but with a session timeout and no custom vocabulary, so component and epic names need constant fixing.
• Wispr Flow — polished cloud dictation with AI formatting; cross-platform but cloud-based, so weigh that against dictating an unreleased roadmap. $144/year.
• Superwhisper — on-device Whisper modes plus optional cloud LLM cleanup; $8.49/month or $249.99 lifetime, no dedicated component-name resolution.

Writing tickets is only part of the day. See our companion guides on how to dictate in Slack for the threads where bugs get reported and how to dictate in Notion for the PRDs behind them. For the architecture comparison, see cloud vs local dictation. Engineers wiring voice into their editor should read how to dictate in Cursor and how to dictate in VS Code.

Basics

Can you dictate in Linear or Jira?
Not natively — neither has built-in speech-to-text for writing issues. Use any system-wide Mac dictation tool to type into the title, description, comments, and sub-tasks of both apps and the web.

Does Linear or Jira have built-in voice input?
No. Linear's AI generates descriptions and reads updates as audio; Jira's Atlassian Intelligence and Rovo generate and rewrite content, and Rovo Desktop dictation only writes AI prompts. None of that transcribes your voice into an issue field.

Setup

How do I set up voice ticket-writing?
Install a system-wide on-device tool, grant Accessibility and Microphone permissions, set a hold-to-talk hotkey, and add your product, component, and epic names to its custom vocabulary. Then click into any field and speak.

How do I get internal names to transcribe correctly?
Register them in your custom vocabulary once. Voibe's custom vocabulary influences the transcription itself, so "CheckoutService" and "Aurora epic" come out right instead of being guessed phonetically.

Workflow

How do I dictate acceptance criteria?
Speak them in Given/When/Then form, one criterion per burst: "Given an empty cart, when I open checkout, then I see an empty-state message." The formulaic structure is voice-friendly and keeps each dictation chunk short.

What's the fastest way to write a full ticket by voice?
Use the Voice Ticket Template: dictate the title, then the As a / I want / so that user story, then each Given / When / Then criterion, then notes — four short bursts in a fixed order, tabbing between fields.

Ticket writing is prose with a fixed shape — a title, a user story, a few acceptance criteria, some notes — which makes it one of the best places to swap typing for voice. Neither Linear nor Jira gives you native dictation, so a system-wide on-device tool is what turns spoken tickets into filled fields, in both apps and the web, with your component names spelled right.

Voibe is the Mac-native, on-device option built for that: download it free (7-day trial, no account), add your product and component names to the custom vocabulary, and dictate your next bug report — and the Slack thread and PRD around it — with the same hotkey.

Keep going:

• How to dictate in Slack — the threads where bugs get reported
• How to dictate in Notion — the PRDs behind your tickets
• How to voice-prompt ChatGPT, Claude, and Cursor — the Five-Part Voice Prompt framework
• Best dictation software for developers — the full buyer's view
• Getting started with Voibe — complete setup guide