How to connect a Zoom MCP with Codex (4 steps)

When you hand Codex a task to build a recording-archival job or a transcript processor, you usually describe the Zoom data in the prompt.

That description skips the details that decide whether the code works. One recording returns several files of different types, the transcript arrives as VTT with a specific structure, and meeting UUIDs follow an encoding rule that breaks endpoint paths.

As a result, Codex generates a download job that handles only the video file, a parser that chokes on the VTT format, or API calls that 404 on a slash-prefixed UUID.

To give Codex direct access to Zoom as it works through your coding tasks, we'll show you how to connect Zoom with Merge Agent Handler's Zoom MCP server.

How it works

Merge Agent Handler connects Codex to the Zoom API through the Merge CLI. You install the CLI, authenticate once, and run a single setup command from your project root.

That command writes a Merge CLI section to your project's AGENTS.md file, which tells Codex when to call merge search-tools and merge execute-tool to reach Zoom.

Once connected, Merge handles OAuth token storage and refresh on your behalf, so you never embed Zoom credentials in your repo or manage per-user authorization state yourself.

Related: How to use the Zoom MCP in Claude Code

Prerequisites

Before getting started, you'll need the following:

• A Merge Agent Handler account
• Codex access (available via the OpenAI platform)
• pipx installed (run pipx --version to confirm, or install via pip install pipx)
• A Zoom account, with cloud recording enabled if you want recording and transcript access

If you want to connect Merge Agent Handler's Zoom MCP with internal or customer-facing agentic products, you can follow the steps in our docs.

1. Install the Merge CLI

Install the Merge CLI with pipx: pipx install merge-api

Verify the install: merge --version

2. Log in to Merge

Run the following to authenticate the CLI with your Merge Agent Handler account: merge login

This links the CLI to your Merge account and stores your session credentials locally.

3. Add Agent Handler to Codex

From the root of the project where you want Codex to reach Zoom, run:

merge setup agents-md

This writes a Merge CLI section to your project's AGENTS.md file so Codex knows to use the CLI when a task needs Zoom data. The command is idempotent, safe to re-run if you need to reset the configuration.

Commit the updated AGENTS.md so the configuration travels with the repo.

Related: A guide to integrating the Zoom MCP with Cursor

4. Authenticate Zoom

Create a Codex task that needs live Zoom data. For example: "Read a recent meeting's recording files and transcript, then scaffold a job that archives the recordings to S3 and stores the transcript text alongside them."

The first time Codex invokes a Zoom tool, a Magic Link will appear to complete connector authentication.

Once authenticated, Codex can reach your Zoom account through Merge for every later task in this project.

{{this-blog-only-cta}}

Zoom MCP FAQ

In case you have more questions on setting up and using the Zoom MCP with Codex, we've addressed several more commonly-asked questions below.

What can you do once the Zoom MCP is connected to Codex?

With Zoom connected, Codex can:

• Read a real recording's file list before scaffolding an archival job: pull a live recording so the job it generates handles every file type the meeting returns, video, audio, chat, and transcript, rather than just the one a prompt mentioned

• Pull a real VTT transcript before generating a summarization feature: fetch an actual transcript so the parsing and chunking code it writes matches the real timestamp and speaker format before it feeds the text to a model

• Read a meeting object before scaffolding meeting-creation code: fetch a live meeting so the settings payload it generates uses fields the API actually returns

• Pull participant data before generating an attendance-report feature: fetch a meeting's participant list so the duration and attendance logic it writes is built on the real join and leave timestamp format

• Read recording metadata before generating test fixtures for a sync job: fetch real recordings so the fixtures and assertions it produces cover the file-list and UUID cases the sync will hit in production

Why use Merge Agent Handler vs. a self-hosted Zoom MCP server?

You can build your own MCP server on top of Zoom's API. For one developer on one account, it's workable: create an OAuth app, run the consent flow, and write tool schemas for the endpoints your tasks touch.

It gets harder once more than one person uses it or tasks run in CI.

Zoom issues short-lived access tokens with refresh tokens, and every connected user needs their own token pair, refresh cycle, and handling for revoked access. A silent refresh failure means a Codex task quietly runs against stale or missing data, with no central view of which connections are live.

A self-hosted server doesn't change that. You still own the OAuth plumbing and the app, and now you maintain the server too. Community Zoom MCP builds vary in endpoint coverage and upkeep, so you're betting on someone else's maintenance.

Merge Agent Handler handles credential storage and token refresh across every connected user. You can scope exactly which Zoom operations a Codex task can call, and every call is logged with identity, timestamp, and inputs.

For an agent reading recordings and transcripts that contain meeting content, scoped access plus full audit logging is the foundation you want in place first.

Why connect Zoom to Codex?

Zoom holds the recording file lists, transcripts, participant data, and meeting objects that meeting-integration code has to handle exactly.

Codex tasks that build archival jobs, transcript processors, or meeting automations need that ground truth to produce code that works against real data.

The alternative is describing the data in the prompt, and those descriptions skip the hard parts. A summary rarely mentions that a recording returns several file types, that the transcript is VTT with a specific structure, or that meeting UUIDs need double encoding, so Codex generates code that breaks on the first real recording.

Connecting Zoom lets Codex read the actual data when a task needs it. The real recording file list before an archival job, the real VTT before a transcript parser, the real meeting object before creation code: Codex works from the account itself, not a paraphrase of it.

Can I use Merge Agent Handler's Zoom MCP with my employees?

Yes, Agent Handler for Employees is built to help organizations provision, secure, and govern how employees connect AI tools like Codex to systems like Zoom.

Common patterns include:

• Provisioning and access control via SCIM with identity providers like Okta and Microsoft Entra ID, so IT can manage which Zoom accounts and data types an employee's agent can reach by role or team

• DLP and policy enforcement on tool calls, so admins can block queries that would return sensitive recording or transcript content before results reach a Codex task's context

• User-level audit logging so security and IT teams can review which meetings, recordings, and transcripts were accessed, by which employee identity, and when

Taken together, employees can use the Zoom MCP to scaffold archival jobs against real recording file lists, generate transcript processors grounded in the actual VTT structure, and build meeting automations with correct settings payloads, and more, while IT keeps centralized control over which data each agent can reach.