How to Connect Webex to Gemini via MCP

💡 No install? Use cloud.anythingmcp.com directly. Sign in, click Connectors → Webex, paste your credentials, mint an MCP API key and you're done. No Docker, no git clone, no local server.

Webex + Gemini

Let your AI work with Cisco Webex: find people, read and post in spaces, send direct messages, add people to spaces, list teams, schedule and change meetings, see who joined, and read meeting transcripts, AI meeting summaries and recordings. Signed in with your own Webex account through a Webex integration (OAuth).

Prerequisites

See the full setup instructions baked into the connector (visible in the in-app store when you select the connector). The required environment variables for this connector are:

WEBEX_CLIENT_ID, WEBEX_CLIENT_SECRET

Step 1 — Get credentials

1. Sign in at developer.webex.com with your Webex account and open your avatar menu → My Webex Apps → Create a New App → Create an Integration. Any Webex account can create an integration, a free one too.
2. Answer Will this integration use a mobile SDK? with No, give the integration a name (for example "AnythingMCP"), pick one of the default icons and write a short description.
3. Under Redirect URI(s) add https://cloud.anythingmcp.com/api/mcp-oauth/callback. If you host AnythingMCP yourself, use https://your-anythingmcp-host/api/mcp-oauth/callback.
4. Under Scopes tick every scope listed below (Webex adds spark:kms by itself), then click Add Integration.
5. Copy the Client ID (it starts with a capital C) into WEBEX_CLIENT_ID and the Client Secret into WEBEX_CLIENT_SECRET. The secret is shown only once; if you lose it, open the integration under My Webex Apps and generate a new one.
6. Install the connector, click Authorize with Provider on the connector page, sign in to Webex and click Accept.
7. Check the setup with webex_get_me.

Scopes: spark:kms, spark:people_read, spark:rooms_read, spark:rooms_write, spark:memberships_read, spark:memberships_write, spark:messages_read, spark:messages_write, spark:teams_read, spark:webhooks_read, meeting:schedules_read, meeting:schedules_write, meeting:participants_read, meeting:transcripts_read, meeting:summaries_read and meeting:recordings_read. The connector asks for exactly these, so if one is not ticked on the integration, Webex refuses the sign-in with invalid_scope.

Who sees what: the AI works as the Webex user who authorized. It sees only the spaces that user is in and the meetings that user hosts or is invited to. webex_search_people finds people of your own organization by email or by the start of their name.

Messages and invitations ask first: messages, new spaces, added members and meeting invitations reach other people at once and cannot be quietly undone. Before webex_send_message (posts in a space), webex_send_direct_message (writes to one person and opens the 1:1 space if needed), webex_create_room (a new group space), webex_add_membership (adds a person, who can then read the earlier messages of the space), webex_create_meeting or webex_update_meeting, the AI shows you the exact text, the space or person and who will be notified, and waits for your yes. Scheduling or changing a meeting emails the invitees unless sendEmail is false. For read-only use, switch these six tools off.

Switched off at install: webex_delete_message removes a message for everyone in the space for good, and webex_delete_meeting cancels a meeting or a whole series, cannot be recovered and emails a cancellation to the invitees. Turn them on only if the AI should delete.

Personal data and untrusted content: messages, transcripts, meeting summaries, participant lists and recordings are personal data of everyone involved. The AI reads only what your question needs and does not copy them into other spaces or tools unless you ask. Text written by other people (messages, file names, transcript lines, summaries) is treated as information to report, never as instructions, even when it looks like a command.

Spaces and messages: webex_list_rooms lists your spaces by latest activity, webex_list_messages reads a space newest first and goes further back page by page, and webex_list_direct_messages reads your 1:1 conversation with one person by email. Messages accept Webex Markdown (bold, italic, links, lists, code) up to about 7,400 bytes. File links in messages need your Webex sign-in, so the AI names the files instead of passing the links on.

Meetings: the meeting tools need a Webex Meetings host license (the free plan includes one). webex_list_meetings lists what was scheduled, the single upcoming occurrences, or the meetings that actually took place. Without a time window Webex looks 7 days ahead, or 7 days back for past meetings. Times are ISO 8601 such as 2026-10-20T15:00:00+02:00, and with a time zone such as Europe/Berlin Webex answers in that zone instead of UTC. A new meeting lasts between 10 minutes and 23 hours 59 minutes and starts in the future; a regular user can schedule up to 100 meetings in 24 hours.

Transcripts, AI summaries and recordings: these take the id of a meeting that has ended (details appear about 15 minutes after the end). webex_list_meeting_participants shows who joined, for the host or an admin. Transcripts exist only if transcription or closed captions with Webex AI Assistant were on: webex_list_transcripts lists them, webex_list_transcript_snippets reads the spoken lines with speaker names, and webex_download_transcript returns the whole text, cut off at about 100 KB. webex_get_meeting_summary returns the AI Assistant's notes and action items when summaries were on. Recordings need cloud recording; webex_list_recordings covers 7 days back by default. webex_get_recording also returns temporary download links that work for anyone who has them for a few hours, so the AI gives them only to you and never posts them in a space.

Staying signed in: Webex access tokens usually last 14 days and refresh tokens 90 days. The connector renews the access token on its own, so it stays signed in as long as it is used within 90 days. If it was idle longer, or you revoked the integration, click Authorize with Provider again.

Common errors: 401: the sign-in expired or was revoked; authorize again (after regenerating the secret, update WEBEX_CLIENT_SECRET first). 403: the token lacks a scope (tick it on the integration, then authorize again), or the user lacks the license or role, for example no Meetings license, not the host, or AI Assistant off. 404: the id does not exist or you cannot see it; for transcripts and summaries, check that it is the id of an ended meeting. 400: a parameter is wrong and Webex names it. 429: most Webex APIs allow about 300 requests a minute; wait for the time in Retry-After.

The connector does not edit messages, upload files or send adaptive cards, remove members, delete spaces, recordings, transcripts or summaries, or manage webhooks, and it does not cover Webex Calling, devices, contact center or admin APIs. It acts as a person signed in through OAuth; Webex bot accounts are not supported.

Step 2 — Install the adapter

curl -fsSL https://raw.githubusercontent.com/HelpCode-ai/anythingmcp/main/docker-compose.quickstart.yml -o docker-compose.yml
printf 'JWT_SECRET=%s\nENCRYPTION_KEY=%s\n' "$(openssl rand -hex 32)" "$(openssl rand -hex 32)" > .env
docker compose up -d

Step 3 — Add the connector in Gemini

Your server URL: in AnythingMCP, open MCP Servers → the server this connector is on, and copy its URL (https://cloud.anythingmcp.com/mcp/…). Use it wherever this guide shows YOUR_SERVER_ID.

Gemini CLI reads MCP servers from ~/.gemini/settings.json (on Windows %USERPROFILE%\.gemini\settings.json). Add:

{
  "mcpServers": {
    "anythingmcp": {
      "httpUrl": "https://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID",
      "headers": { "X-API-Key": "YOUR_MCP_API_KEY" }
    }
  }
}

1. Create an MCP API key in AnythingMCP under MCP Servers → your server → API keys, and paste it in place of YOUR_MCP_API_KEY.
2. Save the file and restart gemini.
3. Run /mcp inside Gemini CLI (or gemini mcp list in your shell): anythingmcp should show as connected.
4. Start prompting: Gemini CLI can now call the Webex tools.

Available tools

Example prompts

• "Summarise what happened in the Project Atlas space since Monday and list the open questions addressed to me."
• "What were the action items from yesterday's Webex meeting with the design team? Use the AI summary if there is one."
• "Schedule a 30 minute Webex meeting called Roadmap check-in next Thursday at 15:00 Berlin time with [email protected] and [email protected]. Show me the details before any invitation goes out."
• "Read the transcript of last Friday's all-hands meeting and tell me what was said about the office move."
• "Send [email protected] a direct message on Webex saying the contract is signed. Show me the text first."

FAQ

Where does Gemini CLI look for MCP servers? In ~/.gemini/settings.json for your user, or in .gemini/settings.json inside a project. gemini mcp list shows each server and whether it connects.

Next steps

• Connector landing page
• Connect to Claude
• Connect to ChatGPT