Press a key.
Speak. Paste.

Parleq /PAR-lek/ is a local-first AI dictation app for macOS. Hold a global hotkey, talk, and watch your words appear — cleaned up, punctuated, and pasted into whatever app was focused. Audio stays on your Mac.

New · Enterprise SSO for Okta, Entra & any OIDC IdP
Download for macOS v0.20.0 View source

Apache-2.0 · macOS 14+ · Apple Silicon required

The overlay

Talk it into shape.

Most dictation tools clean up your words once and you're stuck with whatever they decide. Parleq lets you keep talking. While the overlay is open, hold the hotkey again — your next sentence becomes an edit instruction over the text you already have. Reshape tone, reach for a one-tap preset, point at a window for context, or teach it a word it keeps missing. When it's right, tap the hotkey to paste.

Everything it does

Eleven capabilities, one quiet hotkey.

On-device speech recognition

Parakeet TDT v3 runs on the Apple Neural Engine — audio never leaves your Mac, and nothing is written to disk.

Compliance-friendly by default

Audio stays in memory, logs are metrics-only, secrets live in the Keychain, and dictation fails closed when you sign out.

Recovery built in

Recent cleaned dictations stay in an in-memory list you can copy or re-paste — wiped on quit, never written to disk.

Teach it the words it gets wrong

A custom dictionary biases the speech model toward your names and jargon, with aliases and a per-term opt-out.

Learns from your corrections

Opt-in: when you keep fixing the same word, Parleq asks your existing cleanup model to suggest dictionary improvements.

One-tap transform presets

Your own named rewrites appear as chips on the overlay — tap to restyle before pasting, or set one as an app default.

Reference Windows

Point at a window, file, or clipboard so the LLM sees your context — captures stay in memory and never touch disk.

Use the AI you already have

Pick Gemini, OpenAI, Bedrock, Vertex, or Azure — with pasted keys or the CLI sign-in your other tools already use.

Pastes where you started

The focused app is captured at hotkey-down, so cleaned text lands in the right window even if your attention drifts.

A home of its own

Hold the hotkey and tap P for the Parleq window — Recent Dictations, Stats, Settings, and About, with no Dock clutter.

MDM-deployable for IT admins

Push a managed profile that pins the provider, model list, OIDC issuer, and destinations so users can't re-point the app.

Want the full pipeline, stage by stage? Read How it works →

For organizations

Built for the enterprise.

One corporate sign-in backs cloud cleanup for every employee — not just the engineers with a CLI configured. The user authenticates once with your identity provider (Okta, Microsoft Entra ID, or any OIDC IdP); Parleq federates that into AWS Bedrock or Google Cloud Vertex AI and mints short-lived per-cloud credentials at call time. No per-user API keys to distribute or rotate, and no long-lived cloud secret on the device.

Get Parleq

  1. 1.
    Download Parleq v0.20.0SHA-256 available alongside for verification.
  2. 2.
    Open the DMG and drag Parleq.app onto the Applications folder.
  3. 3.
    Launch from Applications. Grant Microphone + Accessibility permissions on first launch. Screen Recording is requested lazily the first time you use Reference Windows — not at startup.
  4. 4.
    The setup wizard walks you through picking a cleanup provider (or skipping cleanup entirely). Re-runnable from the menu bar's Run Setup… entry at any time.
  5. 5.
    Hold right Option to dictate. Release to clean and paste. Press ? anytime while dictating for a cheat sheet of every gesture.

First launch downloads the speech model (~150 MB). The menu-bar icon switches from a download glyph to a microphone when ready — typically 30–60 seconds on first run, under 5 seconds on subsequent launches.