Post

Orbit AI Demo

5 min read
#ai #reading-list #openai #google-deepmind #llms #genai

Demo: Orbit with Claude Code (5 min)

Pre-Demo Setup

  1. Local MongoDB running on localhost:27017
  2. Atlas API credentials configured
  3. Claude Code connected to Orbit MCP server
  4. Environment variables set:
    export MONGODB_CONN_LOCAL="mongodb://localhost:27017"
export MONGODB_CONN_ATLAS="mongodb+srv://user:pass@cluster.mongodb.net"  # Set after cluster creation

Demo Flow

Opening (30 sec)

You say: "I'm going to show you Orbit - an AI-powered MongoDB mission control. I can manage both local MongoDB databases and MongoDB Atlas infrastructure using natural language through Claude Code."

Prompt:

List my available MongoDB connections

Expected: Shows local connection is registered

Act 1: Build Local Database (1.5 min)

You say: "Let's start by building a telco customer database locally."

Prompt 1:

Connect to my local MongoDB and create a database called orbit-telco

Prompt 2:

Create a plans collection in orbit-telco with these service plans:

[
  { "_id": "plan_unlimited_pro", "name": "Unlimited Pro", "type": "postpaid", "monthlyRate": 85.00, "dataLimit": null, "features": ["5G", "hotspot_50gb", "hd_streaming"] },
  { "_id": "plan_unlimited_basic", "name": "Unlimited Basic", "type": "postpaid", "monthlyRate": 65.00, "dataLimit": null, "features": ["5G", "hotspot_10gb"] },
  { "_id": "plan_family_share", "name": "Family Share 100GB", "type": "postpaid", "monthlyRate": 120.00, "dataLimit": 102400, "features": ["5G", "multi_line_discount"] },
  { "_id": "plan_prepaid_30", "name": "Prepaid 30", "type": "prepaid", "monthlyRate": 30.00, "dataLimit": 5120, "features": ["4G"] },
  { "_id": "plan_prepaid_50", "name": "Prepaid 50", "type": "prepaid", "monthlyRate": 50.00, "dataLimit": 15360, "features": ["5G"] },
  { "_id": "plan_business_elite", "name": "Business Elite", "type": "postpaid", "monthlyRate": 95.00, "dataLimit": null, "features": ["5G", "priority_support", "international_roaming"] }
]

Prompt 3:

Now create a customers collection with these subscribers:

[
  { "_id": "cust_1001", "firstName": "Maria", "lastName": "Santos", "planId": "plan_unlimited_pro", "region": "northeast", "accountStatus": "active" },
  { "_id": "cust_1002", "firstName": "James", "lastName": "Chen", "planId": "plan_family_share", "region": "west", "accountStatus": "active" },
  { "_id": "cust_1003", "firstName": "Aisha", "lastName": "Patel", "planId": "plan_prepaid_50", "region": "south", "accountStatus": "active" },
  { "_id": "cust_1004", "firstName": "Robert", "lastName": "Kim", "planId": "plan_business_elite", "region": "west", "accountStatus": "active" },
  { "_id": "cust_1005", "firstName": "Sofia", "lastName": "Rodriguez", "planId": "plan_unlimited_basic", "region": "south", "accountStatus": "active" },
  { "_id": "cust_1006", "firstName": "David", "lastName": "Thompson", "planId": "plan_prepaid_30", "region": "midwest", "accountStatus": "active" },
  { "_id": "cust_1007", "firstName": "Emily", "lastName": "Okonkwo", "planId": "plan_unlimited_pro", "region": "northeast", "accountStatus": "active" },
  { "_id": "cust_1008", "firstName": "Michael", "lastName": "Brown", "planId": "plan_business_elite", "region": "northeast", "accountStatus": "active" },
  { "_id": "cust_1009", "firstName": "Lisa", "lastName": "Nguyen", "planId": "plan_family_share", "region": "west", "accountStatus": "active" },
  { "_id": "cust_1010", "firstName": "Carlos", "lastName": "Mendez", "planId": "plan_prepaid_50", "region": "south", "accountStatus": "suspended" }
]

Prompt 4:

Create a billing collection with January charges for these customers:

[
  { "_id": "bill_1001_2024_01", "customerId": "cust_1001", "period": "2024-01", "planCharge": 85.00, "roamingCharge": 0, "taxes": 7.23, "totalDue": 92.23, "status": "paid" },
  { "_id": "bill_1002_2024_01", "customerId": "cust_1002", "period": "2024-01", "planCharge": 120.00, "roamingCharge": 0, "taxes": 10.20, "totalDue": 130.20, "status": "paid" },
  { "_id": "bill_1003_2024_01", "customerId": "cust_1003", "period": "2024-01", "planCharge": 50.00, "roamingCharge": 0, "taxes": 4.25, "totalDue": 54.25, "status": "paid" },
  { "_id": "bill_1004_2024_01", "customerId": "cust_1004", "period": "2024-01", "planCharge": 95.00, "roamingCharge": 85.00, "taxes": 15.30, "totalDue": 195.30, "status": "paid" },
  { "_id": "bill_1005_2024_01", "customerId": "cust_1005", "period": "2024-01", "planCharge": 65.00, "roamingCharge": 0, "taxes": 5.53, "totalDue": 70.53, "status": "paid" },
  { "_id": "bill_1006_2024_01", "customerId": "cust_1006", "period": "2024-01", "planCharge": 30.00, "roamingCharge": 0, "taxes": 2.55, "totalDue": 32.55, "status": "paid" },
  { "_id": "bill_1007_2024_01", "customerId": "cust_1007", "period": "2024-01", "planCharge": 85.00, "roamingCharge": 5.00, "taxes": 7.65, "totalDue": 97.65, "status": "paid" },
  { "_id": "bill_1008_2024_01", "customerId": "cust_1008", "period": "2024-01", "planCharge": 95.00, "roamingCharge": 150.00, "taxes": 20.83, "totalDue": 265.83, "status": "paid" },
  { "_id": "bill_1009_2024_01", "customerId": "cust_1009", "period": "2024-01", "planCharge": 120.00, "roamingCharge": 0, "taxes": 10.20, "totalDue": 130.20, "status": "paid" },
  { "_id": "bill_1010_2024_01", "customerId": "cust_1010", "period": "2024-01", "planCharge": 50.00, "roamingCharge": 0, "taxes": 4.25, "totalDue": 54.25, "status": "overdue" }
]

Act 2: Query Local (45 sec)

You say: "Now let's query our data with natural language."

Prompt 5 (Simple):

Show me all customers on unlimited plans in the orbit-telco database

Prompt 6 (Analytical):

Calculate total revenue by plan type for January in orbit-telco. Join billing with customers and plans to get plan types.

Expected: Shows aggregation with postpaid vs prepaid revenue breakdown

Act 3: Provision Atlas (45 sec)

You say: "Now let's provision a production cluster on MongoDB Atlas."

Prompt 7:

Create an M10 cluster called telco-prod in AWS us-east-1 for my Atlas project

Note: If demo timing is tight, have cluster pre-created and just show:

List my Atlas clusters and their status

Act 4: Migrate to Atlas (1 min)

You say: "With both connections available, I can migrate data between them."

Prompt 8:

Connect to my Atlas cluster and list available connections

Prompt 9:

Copy the plans, customers, and billing collections from local orbit-telco to Atlas orbit-telco database

Prompt 10:

Verify the migration - count documents in each collection on both local and Atlas

Act 5: Atlas Management (45 sec)

You say: "Orbit also gives me full Atlas administrative capabilities."

Prompt 11:

Show my Atlas billing summary for this organization

Prompt 12:

Run a security audit - show all database users and IP access list entries for my project

Closing (15 sec)

You say: "With Orbit, I can manage the full lifecycle - from local development to production Atlas - all through natural language. No CLI commands to memorize, no console navigation. Just describe what you need."

Quick Recovery Prompts

If something goes wrong, use these:

  • "List databases on local connection"
  • "Show collections in orbit-telco"
  • "What Atlas clusters do I have?"
  • "Show my Atlas projects"

Key Points to Emphasize

  1. Multi-connection: Same tool works across local and Atlas
  2. Natural language: No need to know MongoDB query syntax
  3. Full lifecycle: Dev (local) → Production (Atlas) in one tool
  4. Admin + Dev: Both database operations AND Atlas management
  5. Real queries: Aggregations, joins, analytics - not just CRUD

Cleanup

After the demo, remove everything we created:

Prompt (Drop local database):

Drop the orbit-telco database on local MongoDB

Prompt (Drop Atlas database):

Drop the orbit-telco database on Atlas

Prompt (Delete Atlas cluster):

Delete the telco-prod cluster from Atlas

Prompt (Verify cleanup):

List databases on local and list my Atlas clusters to confirm cleanup

Manual Cleanup Commands (if needed)

If Orbit isn't available, use these directly:

# Local MongoDB
mongosh --eval "use orbit-telco" --eval "db.dropDatabase()"

# Atlas cluster (via Atlas CLI or Console)
atlas clusters delete telco-prod --projectId <your-project-id> --force