Vortex
Download

Local-first AI workspace

Meet Vortex.
Your AI, on your machine.

One desktop app for every model you use — and every file stays local.

Download for WindowsMac build coming soon →

Free. Mac & Windows. No account required to start.

See how it works ↓

Works with every major provider

OpenAIAnthropicGoogleDeepSeekMistralKimiOpenRouter

+ any OpenAI-compatible endpoint

Get started

Three steps to your own AI workspace.

01

Install Vortex.

One download. Mac or Windows. No account.

02

Connect a model.

Use the Vortex Gateway in one click — or paste an API key from any provider. Switch any time.

03

Get to work.

Chat, attach local files, run skills, schedule jobs, message Vortex from WhatsApp.

Download for WindowsMac build coming soon →

Why Vortex

A serious tool, built the way serious tools should be built.

Local-first by design

Your chats, files, and memory live in a SQLite database on your own disk. Nothing leaves your machine unless you send it.

How local-first works →

Every model, one app

OpenAI, Anthropic, Google, DeepSeek, Kimi, OpenRouter — or the Vortex Gateway. Switch mid-conversation.

See it switch →

A workspace, not a chat box

MCP servers. Skills. Scheduled jobs. Local file tools. RAG over your own sources. The things a real workflow needs.

Tour the workspace →

Reach you anywhere

A WhatsApp bridge puts Vortex one message away from your phone. Ask from the grocery store. Read the answer on the walk back.

See it on WhatsApp →

Local-first

Your data has an address. It’s on your disk.

Every chat, every memory, every uploaded file lives in a single SQLite database in your user folder. Vortex reads your files directly from disk — they’re not uploaded to anyone’s “knowledge base.”

  • ·~/Library/Application Support/Vortex/vortex.dbon Mac
  • ·%APPDATA%\Vortex\vortex.dbon Windows
  • ·No background telemetry
  • ·No silent sync
  • ·The Vortex Gateway is opt-in. Bring your own key and we never see a token.
Vortex reading a local strategy document and grouping its open risks by severity

Multi-provider

Configure any provider in Vortex — your own keys override the managed gateway

One conversation. Any model.

Mid-chat, switch from GPT to Claude to DeepSeek to Kimi without losing context. Vortex keeps your conversation; the model is just a backend. Use the Vortex Gateway for one-click access to every major provider, or paste your own API key and pay the provider directly. Your call.

  • ·Switch provider per chat or per message
  • ·Bring your own key — Vortex never touches it
  • ·Or use the managed gateway with one toggle
  • ·Works with any OpenAI-compatible endpoint, including local models

Workspace

Built for actual work, not chat-bubble demos.

MCP servers

Plug in any Model Context Protocol server and its tools show up in Vortex automatically. Tools are namespaced per server, so nothing collides.

Skills

A skill is just a folder. Drop it in, flip the toggle, and Vortex can use it. Turn it off without deleting anything.

Scheduled jobs

Cron-style scheduling with a per-job model override. Set Vortex to summarize your inbox at 7am with one model and draft your standup with another.

Schedule a job in Vortex — give it a name, a prompt, a cron schedule, and a model tier

WhatsApp bridge

Your AI, one message away.

Vortex listens on WhatsApp through the official Meta Cloud API. Send a message from your phone and get an answer — with context from your chats, your files, your skills. Same Vortex you use at your desk.

Ask it from the grocery store. Get the summary on the walk back. Pick up the conversation when you sit down.

Vortex
online
what’s on my calendar this afternoon?
Three things — 2pm sync with Maya, 3:30 design review, 5pm dentist. The design review prep doc is in your Drive if you want a quick summary.
yes please, short version
• New nav IA proposal
• Cuts top-level items from 7 → 4
• Open Q: where mobile settings live

Pricing

Free to install. Honest pricing on what you use.

Two ways to power Vortex. Pick either. Switch any time.

Bring your own key

Free forever.

Paste an API key from OpenAI, Anthropic, Google, DeepSeek, OpenRouter, or any OpenAI-compatible endpoint. You pay the provider directly. Vortex never sees your key beyond your own machine.

  • Unlimited messages (subject to your provider’s limits)
  • Every model your provider offers
  • No account required

Vortex Gateway

Free or Gold — $10/month

One toggle for one-click access to every major provider. No juggling API keys.

VTS = Vortex Credits. Spent based on each provider’s token pricing — no markup.

FreeGold
Monthly credits1,000,000 VTS10,000,000 VTS
Every supported model✓✓
Credit rollover—✓ while active
Card requiredNoYes
Cancel any time—✓

FAQ

Questions people ask before downloading.

Yes. The app is free. You pay only for the AI usage itself — either to your provider (bring your own key) or via Vortex Gold ($10/month) if you want our managed gateway.
On your machine. A single SQLite database in ~/Library/Application Support/Vortex/ on Mac or %APPDATA%\Vortex\ on Windows. Files you reference in chat are read directly from your disk.
Only what you explicitly send through the Vortex Gateway, and only if you’ve chosen the gateway. Bring-your-own-key traffic goes directly from your machine to the provider you picked. No background telemetry either way.
Anything from OpenAI, Anthropic, Google, DeepSeek, Mistral, Kimi, OpenRouter, or any OpenAI-compatible endpoint — including local models if you run them yourself.
No. Use the Vortex Gateway and you can start chatting in one click. Bring your own key if you’d rather pay providers directly.
With your own key, you pay the provider; Vortex is just the client. With the Gateway, you pay Vortex (Free tier or Gold) and we handle every provider for you. Gateway usage is measured in VTS — 1,000,000/month free, 10,000,000/month on Gold.
They roll over month to month as long as your subscription is active.
Yes. Cancel from inside the app; your remaining credits stay usable until the end of the billing period.
Yes, if you’ve connected a local model (anything with an OpenAI-compatible endpoint on localhost). Cloud providers obviously require an internet connection.
macOS 11+ (Intel or Apple Silicon) or Windows 10/11. About 300 MB of disk space.
Not today. We may open parts of it later.

Stop renting your AI. Run it.

Vortex is free. Mac and Windows. Your data stays where it belongs.

Download for WindowsMac build coming soon →