IntelliJ Prompts Without Typing Friction

Short answer: click into IntelliJ AI Assistant chat, Junie task input, editor comments, or commit message field, press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux), speak, and press again to insert.

Why IntelliJ Workflows Need Rich Prompt Text

IntelliJ is already fast for code edits, navigation, and inspections. The expensive part is writing high-context prose in many places:

• AI Assistant and Junie task prompts
• JavaDoc/KDoc design intent
• commit bodies that explain refactor trade-offs
• run configuration and debugging notes shared in team docs

Voice makes these fields practical to fill with real detail.

IntelliJ-Native Workflows

AI Assistant prompts with symbol-level scope

Instead of "refactor auth," dictate:
"Refactor AuthTokenService and JwtVerifier only. Keep public method signatures unchanged. Replace duplicated token parsing with shared validator. Add tests for expired, malformed, and unsupported algorithm tokens. Do not modify controller layer."

This aligns with IntelliJ users working by class/symbol boundaries.

Junie task orchestration

For Junie, dictate explicit execution boundaries:
"Generate migration plan for moving retry logic from OrderClient into shared ResiliencePolicy. Output steps first, then code edits. Preserve existing Micrometer metrics names. Include rollback steps."

Junie performs better when the task definition includes sequencing and compatibility constraints.

JavaDoc/KDoc for service contracts

IntelliJ can generate stubs, but not intent. Dictate comments that explain behavior:
"Returns cached profile when cache age is under TTL. Falls back to remote fetch on cache miss. Throws ProfileUnavailableException when remote dependency times out after configured retries."

Then add tags quickly with keyboard shortcuts.

Commit tool window with inspection context

IntelliJ commit UI and inspections are often visible together. Dictate message body while reviewing warnings:
"Extracted duplicate null-check logic into ValidationGuard to reduce branch drift across payment adapters. No API changes. Verified with PaymentAdapterTest and static analysis clean."

This gives reviewers reasoning tied to the IDE context.

HTTP client (.http) and scratch files

Many IntelliJ teams keep API calls in .http files and design notes in scratch files. Dictate request notes:
"Use staging token only. Endpoint rate-limited to 10 RPS. Expected 429 after burst of 20 requests."

These notes make local test assets reusable for teammates.

AI Inputs In IntelliJ IDEA

Prompt surfaces where dictation helps:

• AI Assistant chat/composer
• Junie task input fields
• follow-up refine prompts after generated edits

AICHE inserts prompt text only. IntelliJ AI tools execute generation and edits.

How It Works

1. Open the target text field in IntelliJ.
2. Press ⌃+⌥+R / Ctrl+Alt+R.
3. Speak with concrete classes, paths, constraints, and tests.
4. Press again. AICHE inserts cleaned text at caret.
5. Send prompt or save commit.

Audio is streamed for cloud transcription, processed, and discarded immediately after processing, within 1 second. No persistent audio copy.

FAQ

Plugin required?
No. AICHE is a desktop app with global hotkey insertion.

Code-heavy terms and internal symbols?
Use Software Development profile and Custom Vocabulary for internal identifiers.

Linux Wayland?
Usually yes. Some environments may require permission configuration for global hotkeys.

Try it now: open AI Assistant and dictate a symbol-scoped refactor brief with non-goals and test requirements before sending.