v1.4.1 · Rewrite shortcut fix · Now in beta

The AI command palette
that lives in your menu bar.

Select text anywhere. Hit . Rewrite, summarize, translate, code, or compose — streamed inline by an on-device model or any cloud provider you choose. Vehla never gets in your way and never logs your prompts.

Free 7-day trial One license, 2 Macs macOS 14.6+ · Apple Silicon

Bring your own brain. Or six.

OpenAI Anthropic Claude Google Gemini DeepSeek OpenRouter Ollama Gemma 4 on-device MLX on-device
Everything in one palette

Built for the way you actually use your Mac.

A menu-bar app, a global hotkey, and a floating bubble. Three doors into the same fast palette. No browser tab. No copy-paste loop. No model lock-in.

Global hotkey

Highlight, hit a key, get a rewrite.

Vehla works on top of whatever you're in — Mail, Slack, Notion, VS Code, Figma comments. Open the palette with your shortcut, then use the current clipboard, pasted text, or selected text as input.

Streaming

Words appear as they land.

Floating bubble

An orb you can summon.

Optional. Draggable. Snaps to edges.

Local AI

Run Gemma 4 & MLX entirely on your Mac.

Sensitive draft? Flip to local. Nothing leaves the device.

gemma-4-e2b · 4-bit · ~2.5 GB
On-device · Network: off
Cloud providers

Bring your own key. Switch in one click.

OpenAI
Claude
Gemini
DeepSeek
OpenRouter
Ollama
Snippets

Type / for your own prompts.

Reusable. Templated with {selection}, {clipboard}, and {app}.

/standup-update3-line status
/pr-descriptionsummary + test plan
How it works

From cursor to clipboard in four beats.

01

Copy to clipboard or paste

Copy text from any app, or paste into the palette. Pasted text becomes input, not editable query.

02

Hit the hotkey

The palette opens with your context already loaded — clipboard, selection, and the front app.

03

Pick an action

Rewrite, Summarize, Translate, Explain, Generate Shell Command, or one of your own snippets.

04

Auto-copy result

Output streams in markdown. The final response is auto-copied — formatted (RTF) or plain text.

70+ actions, zero clutter

A palette deep enough to live in.

Curated actions for writing, code, terminal, email, social, productivity, and clipboard. Hide what you don't use. Add your own as snippets.

Aa
Rewrite — Professional
Business-ready, polished tone for emails & docs.
Aa
Rewrite — Friendly
Warm and approachable without losing your point.
Aa
Rewrite — Concise
Cuts hedging and filler. Same meaning, half the words.
Aa
Rewrite — Confident
Direct and assured. Drops apologies and "just"s.
Aa
Rewrite — Casual
Relaxed and conversational, like a text to a friend.
Aa
Rewrite — Executive
Senior leadership voice. TL;DR on top.
Aa
Rewrite — Funny
Witty, playful rewrite without becoming a meme.
Aa
Rewrite — Gen Z
Trendy, internet-native cadence.
Aa
Rewrite — Technical
Precise technical register; preserves terminology.
Aa
Rewrite — Customer Support
Empathetic, blame-free, solution-first.
Summarize
Key points, no preamble.
Shorten
Cut length without losing meaning.
Expand
Flesh out an idea with detail and rhythm.
Fix Grammar
Corrects grammar and spelling. Keeps your voice.
To Bullet Points
Turns prose into a clean, parallel list.
To Markdown
Formats content as proper Markdown.
Explain
Plain-English explanation of anything.
Smart Compose
Continues writing from where you left off.
en
Translate → English
Natural English, not a literal translation.
es
Translate → Spanish
Native Spanish phrasing and idioms.
fr
Translate → French
Properly accented, properly idiomatic.
de
Translate → German
Tone-matched German, formal or casual.
it
Translate → Italian
Native Italian phrasing.
pt
Translate → Portuguese
Brazilian or European, your call.
hi
Translate → Hindi
Devanagari with natural register.
ja
Translate → Japanese
敬語 when context calls for it.
Translate → Chinese (Simplified)
简体中文 with native cadence.
ع
Translate → Arabic
RTL-aware, modern standard Arabic.
{ }
Explain Code
Walks through what a snippet actually does.
{ }
Refactor Code
Improves clarity and idiom without changing behavior.
{ }
Convert Language
Port to another language while preserving intent.
{ }
Debug Code
Finds likely bugs and suggests targeted fixes.
{ }
Generate Tests
Writes unit tests with realistic edge cases.
$_
Generate Shell Command
Natural language → a runnable shell command.
$_
Explain Shell Command
Breaks down flags, pipes, and side effects.
$_
Fix Shell Error
Reads the error and suggests a corrected command.
Smart Reply
Context-aware reply matched to the thread tone.
Accept Meeting
Polite acceptance with a confirmation line.
Decline Meeting
Graceful decline, optionally with alternatives.
Follow-up
Friendly nudge that doesn't feel like a nag.
Summarize Thread
TL;DR plus extracted action items.
Refund Reply
Empathetic, blame-free refund response.
#
Make it a Tweet
≤280 chars with a hook and a payoff.
#
Generate Caption
Instagram / TikTok caption with rhythm.
#
Generate Hooks
Five scroll-stopping openers to choose from.
#
Generate Hashtags
Niche-relevant tags, not the obvious ones.
#
Rewrite for Virality
Adds a hook and a clean payoff for shareability.
⌘C
Clipboard Action
Transforms whatever's on the clipboard.
⌘C
Clean Formatting
Normalizes whitespace, smart-quotes, line breaks.
⌘C
Extract Links
Pulls every URL into a clean list.
Extract To-dos
Finds action items in any block of text.
Calendar Event
Title, date, attendees, and an agenda.
Meeting Notes
TL;DR, decisions, action items, open questions.
Summarize Webpage
Bullets plus a one-line TL;DR.
Privacy is a setting

Local for what's yours. Cloud for everything else.

Vehla doesn't run a server. Cloud requests go straight from your Mac to the provider you chose, using a key stored in macOS Keychain. Local mode keeps everything on-device.

Local AI

On-device

Gemma 4 models run natively on Apple Silicon. In Local AI mode, inference stays on the device — the right path for code, contracts, journals, and anything you wouldn't email to a stranger.

Gemma 4 · E4B · 4-bit ~4.5 GB · higher quality
Gemma 4 · E2B · 4-bit ~2.5 GB · default
Gemma 4 · E4B · 8-bit ~9 GB · highest quality

Cloud AI

BYO key

Six providers, one switch. Keys live in Keychain. Choose the provider and model you want, and Vehla shows the active one right in the palette.

OpenAIgpt-4o-mini · gpt-4.1
Anthropic Claudeclaude-sonnet-4-5
Google Geminigemini-2.5-pro
DeepSeekdeepseek-v3
OpenRouterany model id
Ollamalocalhost:11434
Snippets

Your prompts, behind a slash.

Type / in the palette to filter your library. Snippets template against {selection}, {clipboard}, and {app}.

/brief3 bullets, 1 risk, 1 ask
/commitconventional commit msg from diff
/linkedinhook + payoff, ≤220 words
/apologyno excuses, one concrete next step
Personas & memory

Vehla writes like you, not like an LLM.

Set an active persona for tone, vocabulary, and what to avoid. Add memory notes — your name, your team, your style — and Vehla quietly threads them into every prompt.

Founder voice
Direct, technical, no fluff. Never says "leverage".
Customer-support lead
Empathetic, blame-free, action-first.
Senior engineer
Cites the line, names the bug, ships the fix.
vs. the alternatives

Why a menu-bar palette beats another browser tab.

 
Vehla
Chatbot tab
Inline AI in app
Works in every app
Anywhere on macOS
Only in the tab
Per app
Global hotkey
⌥ Space
No
App-specific
On-device models
Gemma 4, MLX
No
No
Bring your own provider
6 cloud + local
Vendor only
Vendor's pick
Streaming markdown out
Inline + auto-copy
In tab
Per app
No server, no logging
Direct provider calls
Routed via vendor
Routed via vendor
Lifetime price
$29.99 once
~$20/mo forever
Bundled
Pricing

One price. One time. Two Macs.

No subscriptions. No seats. No "AI credits". Bring your own key for cloud, or stay on-device for free with local models.

Free trial

$0/ for 7 days

The whole app, every action, every provider. Local models too. No card required.

  • All 70+ palette actions
  • Local Gemma 4 & MLX models
  • All cloud providers (your keys)
  • Snippets, personas, memory
  • Email support
Start 7-day trial
FAQ

Questions before you download.

What is Vehla, in one sentence?+
Vehla is a macOS menu bar app that puts a universal AI command palette behind a global hotkey, so you can rewrite, summarize, translate, code, and compose inside any app you're already using.
Does Vehla work offline?+
Yes. Local AI mode runs Gemma 4 (E2B and E4B) and MLX models entirely on your Mac. With local mode enabled, no text leaves your device. You can mix and match — local for sensitive content, cloud for everything else.
Which providers does Vehla support?+
Cloud: OpenAI, Anthropic Claude, Google Gemini, DeepSeek, OpenRouter, and Ollama. Local: Gemma 4 (E2B and E4B variants, 4-bit) and any MLX-compatible model. API keys are stored in macOS Keychain — never on a server.
How does the 7-day trial work?+
Install Vehla, enter your email, and the trial starts. No credit card. After 7 days you can buy a lifetime license to keep using the app.
What's the licensing model?+
One lifetime license, $29.99, activates on up to two Macs. Free updates for life. Need a third Mac? Deactivate one in Settings and the slot becomes available immediately.
What macOS versions are supported?+
macOS 14.6 Sonoma and macOS 15 Sequoia on Apple Silicon Macs (M1 / M2 / M3 / M4 / M5).
How does Vehla update itself?+
Vehla periodically checks an HTTPS update.json manifest, downloads the new DMG, verifies its SHA-256 hash before installing, then relaunches.
Is my data private?
Vehla doesn't run a server and doesn't log prompts or outputs. Cloud requests go directly from your Mac to the provider you chose. Local mode never touches the network.
Can I refund?
Yes. If Vehla doesn't work for you within 7 days of purchase, email kailaconsulting@outlook.com and we'll refund you, no questions asked.

Stop tab-switching to write.

Vehla puts a serious AI into the keystroke you already half-pressed. Try it free for 7 days.