Sutra now speaks the same language as AI assistants like Claude and Codex. That means you can connect one of these tools directly to your Sutra account and ask it — in plain English — to build and design real things in your spaces: websites, courses, landing pages, blog posts, emails, forms, and more.
You're not copying and pasting between a chatbot and Sutra. The AI works inside your account, with your permissions, building the actual pages and content — while you stay in control of what goes live.
What this is
This guide covers three things:
What you can design on Sutra with AI — and the possibilities it opens
How to connect your AI tool to Sutra — it takes a couple of minutes, with nothing to install
How to actually work with the AI to design directly on Sutra
1 · What you can design on Sutra with AI
Think of the AI as a hands-on partner who knows how to operate Sutra. Anything you'd normally build by clicking around in Sutra, you can now describe and have built for you. That includes:
Websites & pages — whole multi-page websites and landing pages — with rich page content (headings, text, buttons, images, embedded video), a custom site header and navigation menu, and a footer. The AI can build a draft, hand you a preview link, and only publish when you say so.
Courses & learning experiences — course structures with sections and lessons, including Sutra's signature conversational courses that blend group process with AI. Hand the AI an outline, a transcript, or a pile of notes and have it shape the first version.
Community spaces — spaces for groups, cohorts, and memberships — with discussions, member profiles, and the member data that powers Sutra's Member Map.
Blog & Resources posts — articles for your public Resources/blog — drafted, categorized, tagged, and either published or left as a draft for your review.
Help center / knowledge base — support articles and collections, organized and written for you.
Email & your mailing list — draft broadcasts to your list — written, formatted, and queued — so you review and send.
Forms, surveys & member data — surveys and the custom member/contact fields that let you collect and organize information about your people.
Payments & monetization — pricing plans (one-time, recurring, pay-what-you-can), coupons, and paid-access setup for a space.
Automations — “when this happens, do that” flows — for example, when someone registers, send them a welcome sequence or add them to a space.
Members & invitations — invite people (one at a time or in bulk), manage membership, and keep your contact records tidy.
Media — embed Loom, YouTube, or Vimeo, and add images and files to your pages.
Why this matters
From idea to draft in minutes. Describe what you want and get a real, editable draft in your account — not a mockup you have to rebuild.
No blank page. Point the AI at a recording, a PDF, a stack of notes, or your existing site, and let it produce a first version you can react to.
You stay in control. The AI can build directly, or propose drafts and preview links for you to approve before anything publishes.
One place, one partner. Because everything in Sutra already lives together — website, list, courses, community, payments — the AI can work across all of it in a single conversation, instead of you stitching tools together.
Good to know: the AI acts as you — it can only see and change what your own Sutra account is allowed to. And for fine visual polish, the in-app Sutra editor is still the best place to nudge the last details. Think of AI as the way to get most of it built fast, then refine.
2 · Connect your AI tool to Sutra
The only real requirement: make sure you're signed in to Sutra with an account that has manager / editor access to the spaces you want to build in. The AI inherits your permissions — it can do what you can do, and nothing more.
That's it. The “settings” step is really just: be logged in as someone who's allowed to edit.
Everything connects through a single address — Sutra's MCP (Model Context Protocol) server. MCP is just the standard way modern AI tools plug into outside services. The connection link is:
https://mcp.sutra.co/mcp
Wherever your tool asks for an MCP server, connector, or URL, paste that link. When you connect, a browser window opens and asks you to log in to Sutra and approve access — exactly like signing into any other app. Approve it once, and you're connected.
The same approach works in any MCP-capable tool — Claude and Codex both connect this way. The only thing that differs is where you find the “add a connector” setting.
In Claude (Desktop or web)
Open Settings → Connectors.
Click Add custom connector, paste the link, and confirm.
A browser window opens — log in to Sutra and approve access. Sutra now shows as connected, and it's available in any chat.
In Codex
Tell Codex to add a new MCP server using the link above. Codex opens the same browser window where you sign in to Sutra and approve access — there's nothing else to download. Once you approve, you can start asking Codex to build.
Note: menu names and exact wording shift a little between versions and tools, but the recipe is always the same: find where to add an MCP server or connector to your AI of choice → paste the link → log in through the browser when prompted. Nothing extra to install.
3 · Working with the AI to design on Sutra
Once connected, you just talk to the AI normally. A few habits make it dramatically more effective:
1. Tell it which space to work in. Sutra organizes everything into spaces. Point the AI at the right one — by name, or even better with the space link — so it builds in the right place. For example: “Work in my ‘Morning Light’ space.” If you're not sure, ask it to list your spaces first.
2. Start by having it look around. A great opener is: “Connect to Sutra, list my spaces, and tell me what you can build for me.” This confirms the connection is live and grounds the AI in your actual account.
3. Describe the outcome, not the clicks. You don't need to know how Sutra works under the hood. Describe what you want:
“Build a landing page for my new 8-week course — hero section, what's included, who it's for, and a sign-up button. Save it as a draft and send me a preview link.”
“Turn this workshop transcript into a 4-module conversational course outline.”
“Draft three blog posts from these notes and leave them as drafts in my Resources.”
“Create a $200 plan and a ‘pay what you can’ option for the retreat space.”
“Write a welcome email broadcast for new members — don't send it, just draft it.”
4. Ask for a draft and a preview before publishing. For anything public, ask the AI to build a draft and give you a preview link, then review it in your browser. When you're happy, tell it to publish. This is the safest way to work — you see exactly what's going live.
5. Iterate in plain language. Treat it like a collaborator: “Make the headline warmer,” “Add a testimonials section,” “Swap the button to say ‘Join us,’” “Shorten the intro.” It edits the real page in place.
6. Bring your raw material. You don't have to start from scratch. Share a PDF, paste notes, drop in a recording's transcript, or point to your existing website, and ask the AI to build the first version from it.
A simple first session to try
“Connect to Sutra and list my spaces.”
“In [space name], build a draft landing page for [your offer]. Give me a preview link.”
Open the preview, then: “Tweak X and Y.”
“Looks good — publish it.”
Quick reference
Connection link — https://mcp.sutra.co/mcp
Install anything? — No — nothing to download or install.
API key needed? — No — you log in through the browser with your Sutra account.
What it can touch — Only what your account already has permission to edit.
Safest workflow — Ask for a draft + preview link, review, then publish.
Works with — Claude (Desktop and web) and Codex, plus other MCP-capable tools.
With your AI tool connected, you can build across everything on Sutra — your website, list, courses, community, and payments — just by describing what you want. Start small, ask for a draft, and refine from there.
