Skip to main content
Connect your sandbox to Supabase for database, authentication, and storage capabilities. The AI can then directly interact with your backend.

​
Endpoint

POST /api/seemodo-cloud

​
Authentication

All requests require a Supabase API token in the header: 
-H "X-Supabase-Token: your-supabase-access-token"
Get your access token from: Supabase Dashboard → Account → Access Tokens

​
Actions

The endpoint supports three actions:
ActionDescription
activateConnect a Supabase project to the sandbox
create-projectCreate a new Supabase project
statusCheck connection status

​
Activate Cloud

Connect an existing Supabase project to your sandbox.

​
Request Body

​
action
string
required
Must be "activate"
​
projectRef
string
required
Your Supabase project reference ID (e.g., abcdefghijklmnop).
​
sandboxId
string
required
The ID of the active sandbox.
​
sandboxApiUrl
string
required
The API URL of the sandbox backend.

​
Example Request

curl -X POST http://app.seemodo.ai/api/seemodo-cloud \
  -H "Content-Type: application/json" \
  -H "X-Supabase-Token: sbp_your_token_here" \
  -d '{
    "action": "activate",
    "projectRef": "abcdefghijklmnop",
    "sandboxId": "modal-sandbox-1234",
    "sandboxApiUrl": "https://seemodo--sandbox-backend.modal.run"
  }'

​
Success Response

{
  "success": true,
  "project": {
    "ref": "abcdefghijklmnop",
    "name": "My Project",
    "region": "eu-central-1",
    "endpoint": "https://abcdefghijklmnop.supabase.co"
  },
  "apiKeys": {
    "anon": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9...",
    "service_role": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9..."
  },
  "mcpServerName": "supabase_abcdefghijklmnop",
  "mcpAuthUrl": null,
  "mcpAuthNeeded": false
}

​
What Activation Does

1

Verify Project

Confirms you have access to the Supabase project.
2

Get API Keys

Retrieves anon and service_role keys.
3

Configure MCP

Sets up the Supabase MCP server in Seemodo AI.
4

Install Client

Adds @supabase/supabase-js to the sandbox.
5

Create Config

Generates src/lib/supabase.ts with connection details.
6

Add Rules

Copies Supabase best practice rules for the AI.

​
Create Project

Create a new Supabase project.

​
Request Body

​
action
string
required
Must be "create-project"
​
name
string
required
Name for the new project.
​
organizationId
string
required
Your Supabase organization ID.

​
Example Request

curl -X POST http://app.seemodo.ai/api/seemodo-cloud \
  -H "Content-Type: application/json" \
  -H "X-Supabase-Token: sbp_your_token_here" \
  -d '{
    "action": "create-project",
    "name": "My New App",
    "organizationId": "org-abc123"
  }'

​
Success Response

{
  "success": true,
  "projectRef": "newprojectref123",
  "project": {
    "ref": "newprojectref123",
    "name": "My New App",
    "region": "eu-central-1",
    "endpoint": "https://newprojectref123.supabase.co"
  }
}

​
Check Status

Check if a project is connected and accessible.

​
Request Body (POST)

{
  "action": "status",
  "projectRef": "abcdefghijklmnop"
}

​
Query Parameters (GET)

GET /api/seemodo-cloud?projectRef=abcdefghijklmnop

​
Success Response

{
  "activated": true,
  "project": {
    "ref": "abcdefghijklmnop",
    "name": "My Project",
    "status": "ACTIVE_HEALTHY",
    "region": "eu-central-1"
  }
}

​
Files Created in Sandbox

After activation, these files are created:

​
/src/lib/supabase.ts

import { createClient } from '@supabase/supabase-js';

const supabaseUrl = 'https://abcdefghijklmnop.supabase.co';
const supabaseAnonKey = 'eyJhbGciOiJIUzI1NiIs...';

export const supabase = createClient(supabaseUrl, supabaseAnonKey);

​
/AGENTS.md

Contains instructions for the AI about:
  • Using pnpm (not npm)
  • Supabase configuration details
  • MCP server usage examples
  • Best practices for database, auth, storage

​
/.opencode/opencode.json

Configures the Supabase MCP server for direct database access.

​
Using with AI

After activation, you can ask the AI to: 
Create a users table with id, email, name, and created_at columns.
Add RLS policies so users can only read their own data.

Set up authentication with email/password and Google OAuth.

Create a storage bucket for user avatars with size limits.
The AI uses the MCP server to execute these directly on Supabase.

​
Error Responses

​
Invalid Token

{
  "message": "API token required"
}

​
Project Not Found

{
  "message": "Project not found or access denied"
}

​
Setup Failed

{
  "message": "Failed to set up Supabase in sandbox: Connection refused"
}

​
Flow Diagram