Connect your A.I.
Bring every AI you have. Connect each one, flip it on/off like any quality-of-life switch, and give it a job. Maestro then orchestrates — routing each task to the right brain — while the meter shows your spend in real time.
Your AI Shed free
Connect an AI below and we'll open your free Sign-in Podz — a private little "shed" that just houses your AIs so your subscription can run everywhere. You never build in it; it works quietly in the background. Podz never sees your login.
Live.
👤 Sign in your AI — one link, one code
Pick your AI, click Get my sign-in link, approve it, paste the one code it gives you. Your subscription then runs everywhere. Podz never sees your login.
👉 On that page: click the address bar → Ctrl/⌘ + C to copy it → then just click back onto this tab. I grab it and finish automatically.
(Claude just gives you a short code — paste that in the box below instead.)
Prefer a terminal? Open the advanced Shed shell ↴
🛟 Never get stuck
If one AI hits its rate limit, Maestro keeps you going on the next one down your Backup Ladder. Add a free local model (Ollama) at the bottom and you can work forever.
No AI is connected yet — connect at least one, and add a free fallback (a local Ollama model or a free Gemini account) so Maestro never goes dark.
Free options: a local Ollama / LM Studio model (open-source, runs on your computer) or a free Google Gemini account — either one slots in automatically as your last-resort Maestro.
📊 Limits & Resets
🔗 Your Backup Ladder
Not sure which AI? Ask the Concierge
Tell me what you want to build. I’ll recommend the right AI, walk you through hooking it up, and point you to sign-up & billing if it needs a paid key.
Deep reasoning & long context — a strong planner and coder.
⚡ Connect Anthropic Claude
claude-opus-4-8Fast, versatile, huge ecosystem. Great general fallback.
⚡ Connect OpenAI ChatGPT
gpt-4oBig multimodal + long context at low cost. Good for research.
⚡ Connect Google Gemini
gemini-1.5-proWeb-connected answers with live citations — great for research.
⚡ Connect Perplexity
sonarGrok by xAI — fast, current, good all-rounder.
⚡ Connect xAI Grok
grok-2-latestMeta Muse Code — Meta’s new terminal coding agent on Muse Spark 1.2, going head-to-head with Claude Code & Codex. Endpoint/model may need tuning (experimental).
⚡ Connect Meta Muse Code
Muse-Spark-1.2Strong coding + reasoning at a very low price.
⚡ Connect DeepSeek
deepseek-chatGreat for retrieval/RAG and enterprise search.
⚡ Connect Cohere
command-r-plusFast and budget-friendly — great for quick, cheap answers.
⚡ Connect Mistral
mistral-large-latest⚡ Connect Moonshot Kimi K-3
kimi-k3Run a model on your own machine (Ollama). Private, no per-token cost.
⚡ Connect Local / Ollama
llama3New models drop weekly — plug in any that speaks the OpenAI chat format. Add it here, then connect its key like the others.
➕ Add a model
🔑 Has a subscription CLI? (optional)
If this model ships a first-party CLI you log into with your own subscription (like Claude Code / Codex / Gemini CLI), add it here and the “Use my subscription” route turns on for it too. Leave blank for a plain API-key model.
🖥 Add a local model (no key)
Run a model on your own computer (Ollama, LM Studio, llama.cpp) and point Podz at it. No API key, no per-token cost, private. Needs a capable machine, and your local server must be running.
Your Build Team
Bring your own connected AIs — Maestro leads them and stays the only one who checks work off.
Loading…
Two ways to connect — you choose
Every AI can run on your own API key (metered). Any AI with a first-party CLI — Claude (Claude Code), ChatGPT (Codex), Gemini (Gemini CLI), and any custom model you add that has one — can also run on your subscription through a terminal in your Pod, where Podz never touches your login. Maestro works exactly the same either way; only the metered route shows the money-meter.
Run your own first-party CLI (Claude Code, Codex, Gemini CLI…) in your Pod, logged in with your subscription. Podz never sees or stores your login — and there’s no per-token meter here, because your subscription bills you directly.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
Run your own first-party CLI (Claude Code, Codex, Gemini CLI…) in your Pod, logged in with your subscription. Podz never sees or stores your login — and there’s no per-token meter here, because your subscription bills you directly.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
Run your own first-party CLI (Claude Code, Codex, Gemini CLI…) in your Pod, logged in with your subscription. Podz never sees or stores your login — and there’s no per-token meter here, because your subscription bills you directly.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Bring your own API key. This route spends per-token money, so the AI Meter shows your running usage in real time. Your key stays in a locked server-side vault.
This provider has no subscription terminal — API key only.
Services & Tools
Everything else your Podz plug into — payments, code, data, hosting — connected once and stored in the same locked vault as your AI. Add them here, use them anywhere.
Take real payments. Paste your TEST secret key to build safely.
🔗 Connect Stripe
Save your code with full history. Paste a personal access token.
🔗 Connect GitHub
Where your app remembers things. Paste your Supabase/Postgres URL.
🔗 Connect Database
The computer that serves your site 24/7. Paste your host API token.
🔗 Connect Hosting
Your web address (yourname.com). Paste your registrar API token.
🔗 Connect Domain
Gives your AI a memory of your documents. Paste your Pinecone key.
🔗 Connect Vector DB
Advanced (Pro): bring your own deploy endpoint — we POST your built site to it.
🔗 Connect Custom deploy target
Advanced (Pro): store your AWS credentials to deploy to S3 + CloudFront.
🔗 Connect Amazon Web Services
Advanced (Pro): store your Azure Static Web Apps deploy token.
🔗 Connect Microsoft Azure
Advanced (Pro): store your Firebase Hosting deploy token.
🔗 Connect Google Cloud (Firebase)
🐙 Run everything through Git
Link a GitHub repo and every build & deploy auto-commits and pushes there — one clean commit per build. Uses your GitHub token from the vault above; nothing else to set up.
Checking your Git connection…
⚡ First connect your GitHub token on the service card above — then link a repo here.
🔒 Keys live in a locked server-side vault (chmod 600, gitignored). The meter is your own tallied usage estimated from token counts — providers don't expose account balances to a key, so this is the honest, legal way to track spend.
- Open your own Pod (the button below), then open its Terminal.
- Type
claude loginand finish signing in as yourself in the browser. - Done — you’re on your own Claude Code subscription. Podz never sees your login, and there’s no per-token meter.