// mission control · v0.6.1

Mission control
for every AI.

One desktop cockpit for GPT, Claude, Gemini — 49 cloud providers and your local Ollama models — flying on your own API keys.

Windows · macOS · Linux  —  full features, no card required

  • OpenAI
  • Anthropic
  • Google Gemini
  • Mistral
  • Groq
  • DeepSeek
  • xAI
  • Cohere
  • Perplexity
  • Together AI
  • Fireworks
  • OpenRouter
  • Azure OpenAI
  • AWS Bedrock
  • Ollama (local)
  • Cerebras
  • NVIDIA
  • Databricks
  • Replicate
  • Hugging Face
  • + 29 more
The Nexus deck: a dark cockpit chat window with model picker, composer and live inspector panel
every model, one deck — GPT · Claude · Gemini · your Ollama box
real tools in every reply — web search · fetch · any MCP server
Models and providers catalog Model Builder with behavior sliders and QLoRA fine-tuning Local hosting: Ollama runtime with a 30B model running Console: embedded terminal running Claude Code Usage and cost per model
the catalog49 providers, one deck — add one key per provider and every model lands on the board.
model buildershape a local model's identity in minutes — then truly train it on your data with QLoRA.
local hostingpull a 30B model and run it on your own machine — the cloud never sees it.
the consolelive terminals run Claude Code and any CLI without leaving the deck.
flight instrumentstokens and dollars tracked live — per reply, per model, with automatic fallback.

// 01 · the deck

This is the actual app.

// 02 · instruments

Everything on the board.

Six instruments, one window — each of these is a real screenshot.

Models & providers view: Anthropic cloud models beside local Ollama and Nexus models

49 providers, one deck

Add one API key per provider and every model lands in the same cockpit. Search the whole catalog by provider or model name.

Local model hosting view: machine RAM and GPU readout with Ollama runtime running a 30B model

Local models welcome

Point Nexus at Ollama or LM Studio and your own hardware answers beside the cloud — a 30B model on your GPU, no key needed.

Console view: an embedded terminal running Claude Code inside Nexus

A terminal inside

Live terminals run Claude Code and any CLI side by side without leaving the deck. Chat on one side, ship on the other.

Usage & cost view: calls, token counts and estimated spend per model

Costs on instruments

Live token and dollar readouts per reply, spend tracked per model, automatic fallback when a route fails.

Model Builder view: compose a custom local model with personality, behavior sliders and QLoRA fine-tuning

Build your own model

Shape a local model's identity in minutes, then really train it on your data with QLoRA — your weights, on your disk.

Real tools, every model

Web search and page fetch are built in; wire any MCP server for more. Every model gets hands, not just words — and it all stays on your machine.

// 03 · preflight

Flying in three minutes.

  1. 01

    Download

    One installer for Windows, macOS or Linux. The 14-day trial starts on first launch — no account, no card.

  2. 02

    Add a key — or a local model

    Paste any provider's API key, or point Nexus at Ollama. Keys live in your OS keychain, not in files.

  3. 03

    Fly

    Chat with any model, call tools, watch costs, switch providers mid-conversation. The deck is yours.

// 04 · the math

Stop juggling subscriptions.

A stack of subscriptions

  • $20+/month per assistant
  • One model per app
  • Chats live on their servers
  • Throttled at peak hours
  • No local models

Free tiers & browser tabs

  • Login walls and tab sprawl
  • No tools, no files
  • History scattered everywhere
  • Weakest models only
  • Still no local models

// 05 · boarding

Fly free for 14 days. Then $20 a month.

// 06 · flight manual

Questions, answered.

Where do my API keys live?

In your operating system's keychain — Windows Credential Manager on Windows. Never in plain files, and never sent anywhere except directly to the provider you're calling.

Does Nexus work offline?

With local models through Ollama, yes — chat runs entirely on your machine. Cloud providers need a connection, and your license key itself activates offline.

Which models can I use?

Anything your keys unlock: GPT, Claude, Gemini, Mistral, Groq, DeepSeek and 40+ more providers, plus any model you run locally with Ollama.

What happens when the trial ends?

Chat locks until you add a license key. Nothing is deleted — your conversations, keys and settings stay on your machine and unlock with the key.

How do payments work?

You top up credits ($20–$1000) by sending USDC or USDT to a payment address — no card, no processor. Your key then renews itself for 20 credits a month. Crypto transfers can't be reversed, so all sales are final; that's what the free trial is for.

What if I run out of credits?

Your key keeps working through a 3-day grace period while we email you. After that it pauses — nothing is deleted — and topping up reactivates it automatically within minutes.

Do I need an account?

Not for the app — the trial has no signup and there's no telemetry. Buying a key uses a simple email account so your credits and license have a home you can sign back into.

How do updates arrive?

Nexus checks for new releases and installs updates from inside the app — one click on the status bar when a new version lands.

Take the deck.

Fourteen days, every feature, your keys. See what flying feels like.

Start your free 14-day trial