Let your AI assistant follow your connection through Dubai
Open Layer gives an AI assistant the facts about one flight connection through Dubai International (DXB): gates, boarding times, how much time you have and which way to walk. You decide what is shared. Access is for one journey only: it ends when your onward flight departs, and what you shared is then deleted.
1. Copy this address
https://open-layer.dev/mcp
It is the address of a remote MCP server. Nothing else needs to be set up.
2. Add it to your assistant
When you connect, your browser opens a consent page on this site. It asks for your
inbound and onward flight numbers and the date. A server without live flight data accepts only the
demo codes DEMO-CALM and DEMO-LATE, which play a simulated connection. You
then choose what to share and approve. Each time you connect or reconnect you approve again, and
that starts a new journey.
Menu names are those of November 2026 and can change.
Claude
- On the web or in the desktop app, open Customize, then Connectors.
- Choose Add, then Add custom connector.
- Name it Open Layer, paste the address and continue. Keep the settings Claude detects, then choose Add.
- If the connector shows Connect, choose it and approve on the page that opens.
Open the form in Claude with the address filled in
Add it on the web or in the desktop app; it is then available in the mobile app too. The Free plan allows one custom connector.
ChatGPT
- On the web, go to chatgpt.com/plugins, select the plus button, then Add custom MCP server.
- Enter a name, paste the address (it ends in
/mcp) and choose OAuth. - Read the risk warning, select I understand and want to continue, then Create as a plugin and approve on the consent page.
- Install the plugin from your personal plugins. In a new Work chat, type
@to pick it.
Web only. Whether you can add a custom MCP server depends on your plan and workspace settings. The approval you give while creating the plugin is the journey ChatGPT will follow.
Gemini
- On a computer, go to gemini.google.com and open Settings, then Connected Apps.
- Under Custom apps, choose Add a custom app, paste the address and select Next, then approve on the consent page.
Google offers custom apps to personal accounts in the US for now (18 or over, in English, with Keep Activity on). Once added on the web, it also works in the mobile app.
In Gemini CLI:
gemini mcp add --transport http open-layer https://open-layer.dev/mcp
Then sign in from inside Gemini CLI with /mcp auth open-layer.
Other clients
Claude Code (then type /mcp in a session to sign in):
claude mcp add --transport http open-layer https://open-layer.dev/mcp
VS Code: run the command MCP: Add Server, choose HTTP and paste the address.
Cursor: add this to ~/.cursor/mcp.json:
{
"mcpServers": {
"open-layer": {
"url": "https://open-layer.dev/mcp"
}
}
}
MCP Inspector:
npx @modelcontextprotocol/inspector --server-url https://open-layer.dev/mcp --transport http
In any other MCP client, choose Streamable HTTP as the transport and paste the address. An agent that does not speak MCP can use the same operations over plain HTTP: see openapi.json.
No browser redirect?
If your agent cannot open a browser page for you, get a pairing code: you approve on this site and hand the agent a one-time code, which it exchanges for access to the journey.
3. Optional: the Transit Companion skill
A skill is a short instruction file. This one tells an assistant how to behave during your journey: check before answering, only interrupt you when an update changes what you should do, and speak your language. The server already sends the same rules, so the skill is an extra.
- Claude: open Customize, then Skills, and upload the zip as a new skill.
- Gemini app: open Customize, then Skills, choose Create, then Upload, and select
SKILL.md. - Claude Code, Codex, Gemini CLI and Cursor: unpack the zip into your skills folder
(
~/.claude/skillsfor Claude Code,~/.agents/skillsfor the others), so that the file istransit-companion/SKILL.md.
More
- Stop sharing: end your journey early and delete what you shared.
- llms.txt: a short description of this service for language models.
- openapi.json: the same operations over plain HTTP.