In short
VS Code connects to OmniChannels.ai by adding our remote MCP server, https://mcp.broadnet.me/mcp, to .vscode/mcp.json as an http server. A promptString input with password set asks for your API token and stores it securely, so no key is committed. GitHub Copilot agent mode can then send test messages, list WhatsApp templates and read delivery reports while it helps you write integration code.
What you can do
Write code that matches your templates
Copilot lists your approved WhatsApp templates through MCP and writes a sender with exactly the variables each template needs.
Try asking
“List my approved WhatsApp templates, then write a C# method that sends order_update with the order number and delivery date.”
Build and test a webhook endpoint
Copilot writes an endpoint for delivery statuses and inbound messages, then you trigger a test send to see real payloads arrive.
Try asking
“Write a FastAPI webhook endpoint that stores delivery statuses by message ID and accepts fields that vary by channel, then send a test SMS to 447700900123 so I can watch it arrive.”
Check a delivery before you ship
After your code sends a test message, Copilot reads the delivery report so you can confirm the status your handler recorded.
Try asking
“Check the delivery report for the last message sent to 971500000000 and compare its status with what our database recorded.”
Use the spec, not memory
With our OpenAPI spec in the workspace, Copilot uses the exact paths and fields, including which sends need multipart/form-data.
Try asking
“Using docs/omnichannels-openapi.yaml, write a bulk WhatsApp send helper and point out which fields are sent as multipart/form-data.”
Draft Arabic and English copy
Copilot drafts message text in both languages and shows it to you before any send tool is called.
Try asking
“Draft an appointment reminder SMS in Arabic and English under one segment each and show me before sending a test to 966500000000.”
How it connects
VS Code reads MCP servers from .vscode/mcp.json in your workspace. Our server is added as an http server with an Authorization header, and the token comes from an input of type promptString with password set to true, so VS Code prompts for it once and keeps it out of the file you commit.
In GitHub Copilot agent mode, the tools from enabled MCP servers appear in the tools list, where you can switch each one on or off, and VS Code can ask you to confirm tool calls. Keep confirmation on for anything that sends, since each send is a real, billed message. Configuration fields change between VS Code versions, so check its MCP documentation if a field is not recognised.
- File: .vscode/mcp.json, server type http, URL https://mcp.broadnet.me/mcp
- Header: Authorization: Bearer ${input:omnichannels-token}
- Token input: promptString with password: true
- Enable only the tools this workspace needs in the tools list
Set it up
- 1
Request API access
Request test access and receive an API token. Test sends go only to your own verified numbers.
- 2
Get a temporary key
In the admin portal open Integrations, then MCP Integration, and choose Get temporary key.
Guide: MCP server for AI agents - 3
Create .vscode/mcp.json
Add an omnichannels server of type http with our URL, an Authorization header that uses ${input:omnichannels-token}, and a promptString input with password set to true.
Guide: MCP server for AI agents - 4
Start the server and enter the key
Start the server from VS Code, paste your key when prompted, and check that the tools appear in Copilot agent mode.
- 5
Limit the tools
In the tools list, switch off tools this workspace does not need and keep confirmation on for every send tool.
- 6
Test on your own number
Ask Copilot to send a test message to your verified number, then read the delivery report.
Vendor documentation
VS Code is a trademark of its owner. OmniChannels.ai is not affiliated with or endorsed by VS Code; this page describes a connection built with VS Code's own MCP, HTTP or webhook features.
