Connecting OmniMem to GitHub Copilot
GitHub Copilot in VS Code has supported MCP servers natively since VS Code 1.99. It speaks Streamable HTTP with custom headers, which is all OmniMem needs.
Requirements
- VS Code 1.99 or later
- Copilot Free, Pro or Pro+: MCP works out of the box
- Copilot Business or Enterprise: your organisation admin has to enable the "MCP servers in Copilot" policy
Quick setup
Create or edit .vscode/mcp.json in your project root:
{
"servers": {
"omnimem": {
"type": "http",
"url": "http://127.0.0.1:8765/mcp"
}
}
}
Copilot uses "servers" as the top-level key, not "mcpServers". Easy to miss.
If you're running the desktop app, Copy MCP URL in the tray or menu bar menu gives you the address.
If OmniMem has an access token set (MCP_AUTH_TOKEN, or Access token on the Configuration page), use an input variable so VS Code prompts for it and stores it securely:
{
"servers": {
"omnimem": {
"type": "http",
"url": "http://127.0.0.1:8765/mcp",
"headers": {
"Authorization": "Bearer ${input:omnimem_token}"
}
}
},
"inputs": [
{
"type": "promptString",
"id": "omnimem_token",
"description": "OmniMem access token",
"password": true
}
]
}
VS Code asks once and remembers. You can also reference an environment variable directly with ${env:OMNIMEM_TOKEN}.
Coming from 6.x, change "type": "sse" to "type": "http" and /sse to /mcp.
Global vs project config
| Scope | How to get there | Use when |
|---|---|---|
| Workspace | .vscode/mcp.json in the project root |
Share the config with your team through version control |
| User (global) | Command palette: MCP: Open User Configuration |
OmniMem should be available in every project |
Don't configure the same server in both, it causes conflicts.
Agent mode is required
MCP tools are only available in Copilot's Agent mode, not Ask or Chat. To use OmniMem:
- Open Copilot Chat (Ctrl/Cmd + Shift + I)
- Pick Agent from the mode dropdown
- The OmniMem tools are there
In Agent mode Copilot can call OmniMem on its own as part of a multi-step task.
Using OmniMem with Copilot
- "Call the OmniMem briefing tool for this project"
- "Remember this architectural decision in OmniMem"
- "Check OmniMem for anything we've already solved here"
Verifying the connection
VS Code asks you to trust a new MCP server the first time it starts it. Approve it, then ask in Agent mode:
Can you call the OmniMem health tool?
You should get back the record and vector counts per namespace, whether the embedding model is loaded, and the uptime.
Known quirks
- The 128-tool cap: VS Code allows 128 tools across all active MCP servers per request. OmniMem's 48 fit comfortably, but it adds up if you've got a lot of servers. Use the "Configure Tools" icon in the chat panel to switch individual tools off.
- Agent mode only: MCP tools don't work in Chat or Ask mode.
- No MCP resources: Copilot only supports tools. That doesn't matter here, OmniMem only exposes tools.
- Server naming: keep server names in
mcp.jsonfree of whitespace and special characters. - Claude Desktop discovery: VS Code can pick up MCP servers from Claude Desktop. Turn it on with
"chat.mcp.discovery.enabled": truein VS Code settings.
Notes
- If OmniMem isn't answering,
curl http://127.0.0.1:8765/healthzshould return{"status": "ok"}, and the desktop app's tray status line says whether it's running - For a server on another machine, see remote access