Gumloop MCP Server & CLI

A free, open source Gumloop MCP server and CLI that lets Claude Code, Codex, Cursor and other AI agents run your flows, work with your agents and search your Company Brain.

githubStar on GitHubnpmnpm
Navid Moazzezby Navid Moazzez·Updated Oct 5, 2026·16 min read·
MCPCLI
Your computer
macOSWindowsLinux
  1. 1

    Add it with a single command in a terminal.

    Put your own API key and user ID in place of the examples.

    claude mcp add gumloop \
      -e GUMLOOP_API_KEY=YOUR_API_KEY \
      -e GUMLOOP_USER_ID=YOUR_USER_ID \
      -- npx -y @thenavidm/gumloop-mcp-cli@latest
  2. 2

    In Claude Code, type /mcp. Gumloop should show as connected.

    /mcp

Needs Node 22 or newer and a Gumloop API key, on the Pro plan or above. Get it in two minutes.

It's free and open source. If it saves you time, star it on GitHub, so more people find it.

Rate this tool

This free Gumloop MCP server and CLI gives your AI real access to your Gumloop account: flows, agents, sessions, your Company Brain, skills and connected MCP servers. It runs your saved flows, talks to your agents and answers their approvals, searches your Company Brain, manages skills and evaluations, and handles team admin.

It's one install with two ways in. Claude, Codex, Cursor or any other MCP app calls its 92 tools for you, and the same tools work as a CLI that agents like Claude Code, Codex and OpenCode run, or that you type yourself.

In this guide, I'll show you:

  • What the Gumloop MCP server & CLI is, and what you can ask it
  • How to install it and check that it works
  • When to use the MCP server, and when the CLI
  • Every tool, and how to keep your account safe

Every tool on this page comes from the server itself, and every cost was measured in Claude Code.

key_takeaways.mdTL;DR

Key takeaways

It gives Claude, Codex, Cursor and any other MCP app 92 tools for Gumloop: flows, agents and their sessions, your Company Brain, skills, connected MCP servers, evaluations and team admin.
Gumloop's API needs the Pro plan or above, and your key is on the Connectors page.
Every run, message, upload and change waits for your approval, because runs spend credits and can reach your connected apps.
The same tools run as a CLI, which costs your AI nothing until it runs.

What is the Gumloop MCP server & CLI?

The Gumloop MCP server & CLI is a free, open source program that lets AI agents run flows and agents in your Gumloop account for you, in two ways. The MCP server is what an AI app like Claude, Codex or Cursor connects to, through MCP (Model Context Protocol), the open standard AI apps use to call outside tools.

You ask in plain language. Your AI picks the right tool, and the server makes the call to Gumloop's own API.

The CLI is the same program as commands. gumloop-cli list-agents runs the same code your AI runs when you ask which agents you have, whether an agent like Claude Code runs it or you do.

In an AI OS, the layer that connects your AI to your context, your data and your tools, a server like this is how your AI reaches Gumloop. I explain how the parts fit together in how to build an AI operating system, and here's how I use AI myself.

What can you ask it?

Once it's set up, you ask the way you'd ask an assistant. These are real prompts it handles:

Try asking
Which flows do I have, and what inputs does my lead-scraper flow need?
Run my weekly report flow for last week.
Ask my research agent to find five competitors, and show me what it needs approved.
Search our Company Brain for the refund policy.
How did my support agent do this month, by grade?
Which MCP servers does my sales agent use?

The third one shows why this is useful. Your AI starts the agent's session and comes back with exactly what the agent wants to do, so you approve each step yourself.

It works well next to my Google Workspace MCP server and Kit MCP server, so your AI can work in Google Sheets and draft emails to your list in the same chat. Every server I've made is on my MCP servers page.

For your prompts and agents, my free AI agent readiness checker and skill generator help too.

How to install the Gumloop MCP server

Pick your app in the box at the top of this page. Each one is a single command or a pasted block, and they all run through npx, so there's nothing to install first.

Windsurf, Zed and Cline take the same server, though they don't have a tab in the box yet. Follow the box's Cursor tab, and put its block here instead:

AppWhere the block goes
WindsurfIn the Cascade panel, open the Actions menu (…) and choose Open MCP config file, then paste the block.
ClineClick the MCP Servers icon in the Cline panel, open the Configure tab, click Configure MCP Servers, and paste the block.
ZedGo to Settings, then AI, then MCP Servers, and choose Add Server, then Add Local Server. Enter the block's command, arguments and variables, and Zed saves them under context_servers.
Before you start0/3

Watch out: Gumloop has personal keys and team keys. A personal key acts as you, and a team key can act as any team member, so pick the one your work needs.

Set up your Gumloop account

It takes two minutes.

Get your API key0/4

For team flows, also set GUMLOOP_TEAM_ID to your team's ID. Gumloop's older flow tools call it the project ID.

To keep the key out of your app's config, save it alone in a file only you can read, and point GUMLOOP_TOKEN_FILE at that file's full path instead.

Check that it works

Run the doctor with --network. It checks your settings, then reads the agents you can use, to prove the key works. It never starts a run.

npx -y @thenavidm/gumloop-mcp-cli doctor --network

When every line has a check mark, restart your app and ask it which flows you have. A 401 or 403 usually means the account isn't on Pro, or the key was deleted.

In Claude Code, /mcp also shows whether the Gumloop server is connected. It's one of the Claude Code commands worth knowing.

Use the Gumloop CLI

The CLI is the same 92 tools as commands. AI agents that run commands, like Claude Code, Codex and OpenCode, use it on their own, and you can type the same commands in a terminal or a script.

Every tool name becomes a command with dashes, so list_flows runs as gumloop-cli list-flows.

npm install -g @thenavidm/gumloop-mcp-cli
gumloop-cli
gumloop-cli list-flows
gumloop-cli get-input-schema --saved-item-id YOUR_FLOW_ID
gumloop-cli start-flow --payload-file ~/gumloop/weekly-report.json --confirm
gumloop-cli get-run-details --run-id YOUR_RUN_ID

The bare gumloop-cli lists every command, gumloop-cli which <words> finds the command for a task, and gumloop-cli <command> --help shows what a command takes. Every run, message, upload and change needs --confirm, the terminal's version of your approval.

gumloop-cli install <app> adds the server to Claude Code, Codex, Claude Desktop, Cursor, VS Code or Gemini CLI, in each app's own format.

These are the flags you'll use most:

FlagWhat it does
--jsonPrints JSON
--compactPrints the JSON on one line
--select a,b.cKeeps only those fields, and a dotted path goes deeper
--agentMachine mode: compact JSON, no prompts, and it never confirms anything
--confirmLets a write that asks first go ahead
--dry-runShows what a write would do, without doing it

A script can branch on the exit code:

Exit codeWhat it means
0It worked
1Something unexpected went wrong
2The command was typed wrong, or a write needed --confirm
3It wasn't found
4Gumloop rejected the credentials
5Gumloop failed or couldn't be reached
7You hit a rate limit, so wait and try again
10Nothing is set up yet

MCP server or CLI: which one?

Both are the same program with the same 92 tools. The difference is when your AI pays for them, in the tokens it reads.

Every message, in an app that loads every tool
MCP server
56,700 tokens
CLI
Nothing
Every message, in Claude Code
MCP server
1,500 tokens
CLI
Nothing
When Gumloop comes up
MCP server
Nothing more, or in Claude Code the tools it picks
CLI
3,200 tokens, once
20 messages with Gumloop in one, every tool loaded
MCP server
1,134,000 tokens
CLI
3,200 tokens
Works in Claude Desktop's chat
MCP server
Yes
CLI
No, there's no terminal there
Works in a script, a cron job or CI
MCP server
No
CLI
Yes

An app that loads every tool up front sends the full list with every message, whether you mention Gumloop or not. Claude Code doesn't by default: its tool search sends only the tool names and the server's instructions, and loads a tool's full definition when your AI reaches for it.

The CLI costs nothing until Gumloop comes up. Then your agent reads its skill file once and runs the commands it needs. With the skill added, Claude Code also lists its one-line description with every message, about 60 tokens.

When the whole conversation is about Gumloop, the gap closes. Then the MCP server is the better experience, because you ask in plain language and your AI never has to look up a command.

How to spend less

In Claude Code, leave tool search on, which it is unless you've set ENABLE_TOOL_SEARCH=false.

Turn the server off when you aren't working with Gumloop. In Claude Code that's the /mcp panel, and every AI app has its own switch. GUMLOOP_READ_ONLY=1 cuts it to the 50 reading tools.

Or install the CLI and connect the server later, on the days it earns its place. Every tool stays in reach, and the standing cost drops to nothing.

Every number here was measured on October 5, 2026 in Claude Code 2.1.286 with Claude Opus 5.5. It's the same one-line prompt with and without the server connected, once with tool search off so every tool loads, and once with Claude Code's default. The skill was measured the same way, and other apps and models count tokens a little differently.

Flows, agents and approvals

A flow run starts with start_flow, using the inputs get_input_schema lists, and comes back with a run ID. Your AI checks that ID with get_run_details instead of starting the flow again. Only a flow's output steps show up as its final outputs.

An agent works in sessions. create_session can start the agent working right away, and when it needs your OK for a tool, the session pauses until resolve_session_approvals answers. Your AI shows you what the agent wants first. A busy agent takes new messages through its queue.

Runs, chats and Brain searches spend credits, and agents can act in the apps you've connected to Gumloop. That's why each of those asks first, even when the step sounds small.

Team admin tools, like adding people to a workspace or setting a role's credit limit, need the right role in your organization. Full data exports are for Enterprise.

Watch out: Gumloop limits how many agents run at once across your organization: 25 on Pro and 100 on Enterprise. On Pro, extra work is refused rather than queued.

Every Gumloop tool

The tables below list all 92 tools, grouped by what they're for. Of these, 50 only read, and 42 spend credits or change something in your Gumloop account or on your computer.

Flows, runs and files

start_flow
What it does
Runs a saved flow, which can spend credits.
Kind
Asks first
kill_flow
What it does
Stops a flow run and its subflows.
Kind
Asks first
get_run_details
What it does
Reads a run's status and its final outputs.
Kind
Reads
list_workbooks
What it does
Lists your workbooks and their saved flows.
Kind
Reads
list_flows
What it does
Lists your saved flows.
Kind
Reads
get_input_schema
What it does
Shows the inputs a saved flow takes.
Kind
Reads
get_run_history
What it does
Lists a flow's 10 most recent runs.
Kind
Reads
download_file
What it does
Saves a file from a run to a new private file.
Kind
Reads
download_files
What it does
Saves several files from runs to a new private file.
Kind
Reads
upload_file
What it does
Uploads a file, up to 3 MB, for your flows to use.
Kind
Asks first
upload_files
What it does
Uploads several files for your flows to use.
Kind
Asks first

Agents

list_agents
What it does
Lists the agents you can use.
Kind
Reads
create_agent
What it does
Creates an agent.
Kind
Asks first
retrieve_agent
What it does
Reads one agent.
Kind
Reads
update_agent
What it does
Changes an agent's settings.
Kind
Asks first
list_agent_versions
What it does
Lists an agent's saved versions, newest first.
Kind
Reads
retrieve_agent_version
What it does
Reads one version of an agent, with what changed.
Kind
Reads
update_agent_skills
What it does
Adds or removes skills on an agent.
Kind
Asks first
list_agent_mcp_servers
What it does
Lists the MCP servers attached to an agent.
Kind
Reads
attach_agent_mcp_server
What it does
Attaches an MCP server to an agent, or updates it.
Kind
Asks first
detach_agent_mcp_server
What it does
Detaches an MCP server from an agent.
Kind
Asks first
list_artifacts
What it does
Lists the files an agent made.
Kind
Reads
download_artifact_file
What it does
Saves a download link for an agent's file to a new private file.
Kind
Reads

Agent sessions

list_sessions
What it does
Lists an agent's sessions.
Kind
Reads
create_session
What it does
Starts a session with an agent, which can begin work at once.
Kind
Asks first
retrieve_session
What it does
Reads a session, with its messages and state.
Kind
Reads
rename_session
What it does
Renames a session.
Kind
Asks first
send_message
What it does
Sends a message in a session and lets the agent continue.
Kind
Asks first
cancel_session
What it does
Stops a running session.
Kind
Asks first
upload_session_file
What it does
Uploads a file into a session.
Kind
Asks first
resolve_session_approvals
What it does
Answers what a paused session is waiting on, like a tool approval.
Kind
Asks first
list_queued_messages
What it does
Lists the messages waiting in a session's queue.
Kind
Reads
queue_session_message
What it does
Queues a message for after the agent's current turn.
Kind
Asks first
update_queued_message
What it does
Changes a queued message.
Kind
Asks first
delete_queued_message
What it does
Removes a queued message.
Kind
Asks first
send_queued_message
What it does
Sends a queued message now, interrupting the agent.
Kind
Asks first

Models and chat

create_chat_completion
What it does
Sends a chat to any model Gumloop supports, which spends credits.
Kind
Asks first
list_models
What it does
Lists the models you can use.
Kind
Reads
route_model
What it does
Asks Gumloop's router which model it would pick for a message, and why.
Kind
Reads

Connected MCP servers

list_mcp_servers
What it does
Lists the MCP servers you can see.
Kind
Reads
retrieve_mcp_server
What it does
Reads one MCP server.
Kind
Reads
list_mcp_server_tools
What it does
Lists an MCP server's tools.
Kind
Reads
list_mcp_server_resources
What it does
Lists an MCP server's resources.
Kind
Reads
read_mcp_server_resource
What it does
Reads one resource from an MCP server.
Kind
Reads
list_mcp_server_prompts
What it does
Lists an MCP server's prompts.
Kind
Reads
get_mcp_server_prompt
What it does
Fills in one of an MCP server's prompts.
Kind
Reads
call_mcp_tools
What it does
Runs up to five MCP tool calls at once.
Kind
Asks first

Company Brain

search_brain
What it does
Searches your Company Brain, which spends credits.
Kind
Asks first
list_brain_sources
What it does
Lists your Company Brain sources.
Kind
Reads
create_brain_source
What it does
Creates a file-upload source in your Company Brain.
Kind
Asks first
get_brain_source
What it does
Reads one Brain source.
Kind
Reads
delete_brain_source
What it does
Deletes a Brain source and all its files, for good.
Kind
Asks first
list_brain_files
What it does
Lists a source's files and how far indexing has got.
Kind
Reads
upload_brain_files
What it does
Uploads up to 25 files to a Brain source.
Kind
Asks first
delete_brain_file
What it does
Removes a file from a Brain source.
Kind
Asks first
get_brain_source_estimate
What it does
Reads what indexing a source will cost in credits.
Kind
Reads
approve_brain_source
What it does
Approves a draft source, which starts indexing and spends credits.
Kind
Asks first

Skills and browser profiles

list_skills
What it does
Lists your skills.
Kind
Reads
create_skill
What it does
Uploads a skill package with its SKILL.md.
Kind
Asks first
update_skill
What it does
Replaces a skill's files.
Kind
Asks first
delete_skill
What it does
Deletes a skill.
Kind
Asks first
download_skill_file
What it does
Saves a download link for a skill to a new private file.
Kind
Reads
list_browser_profiles
What it does
Lists your browser profiles, the sign-ins an agent's browser uses.
Kind
Reads
import_browser_profile_cookies
What it does
Adds sign-in cookies you provide to a browser profile.
Kind
Asks first

Evaluations

list_evaluations
What it does
Lists an agent's evaluation results.
Kind
Reads
run_evaluations
What it does
Grades up to 200 of an agent's finished sessions.
Kind
Asks first
get_evaluation_metrics
What it does
Counts an agent's grades and tags over a time window.
Kind
Reads
retrieve_evaluation
What it does
Reads one evaluation result.
Kind
Reads
get_evaluation_config
What it does
Reads an agent's evaluation settings.
Kind
Reads
update_evaluation_config
What it does
Changes an agent's evaluation settings.
Kind
Asks first
get_evaluation_options
What it does
Lists the values evaluation settings can take.
Kind
Reads
list_organization_evaluations
What it does
Lists your organization's evaluations.
Kind
Reads
create_organization_evaluation
What it does
Creates an organization evaluation.
Kind
Asks first
get_organization_evaluation
What it does
Reads one organization evaluation.
Kind
Reads
update_organization_evaluation
What it does
Changes an organization evaluation.
Kind
Asks first
delete_organization_evaluation
What it does
Deletes an organization evaluation.
Kind
Asks first
set_organization_evaluation_targets
What it does
Sets which agents an evaluation grades.
Kind
Asks first
run_organization_evaluation
What it does
Grades up to 200 sessions with an organization evaluation.
Kind
Asks first
list_organization_evaluation_results
What it does
Lists an evaluation's results across your agents.
Kind
Reads
get_organization_evaluation_result
What it does
Reads one evaluation result in full.
Kind
Reads
get_organization_evaluation_metrics
What it does
Counts an evaluation's grades over a time window.
Kind
Reads

Teams and organization admin

get_organization_audit_logs
What it does
Reads your organization's audit log for a time period.
Kind
Reads
manage_project_users
What it does
Adds or removes people from a workspace.
Kind
Asks first
manage_permission_group_users
What it does
Adds or removes people from a custom role.
Kind
Asks first
list_role_credit_limits
What it does
Lists each custom role's monthly credit limit.
Kind
Reads
get_role_credit_limit
What it does
Reads one role's monthly credit limit.
Kind
Reads
set_role_credit_limit
What it does
Sets or clears a role's monthly credit limit.
Kind
Asks first
export_data
What it does
Starts a full data export for your organization, on Enterprise.
Kind
Asks first
get_export_status
What it does
Reads an export's status, and saves the file once it's ready.
Kind
Reads
list_teams
What it does
Lists your teams.
Kind
Reads
list_organizations
What it does
Reads your organization.
Kind
Reads

Your accounts

list_accounts
What it does
Lists the accounts you've set up, without their keys.
Kind
Reads

Is the Gumloop MCP server safe?

Runs spend credits, and agents can act in your connected apps. So 42 tools ask first: every flow run, chat, session message and approval, every upload, every change to agents, skills, Brain sources and evaluations, and every admin change. They're the ones marked "Asks first" in the tool tables above.

Reading flows, sessions, results and settings never asks, because it changes nothing.

When your AI calls one of them, the call needs a yes first. Where you use it decides who gives it:

  • Claude Code (2.1.246 and later) asks you before every call, in every one of its permission modes.
  • An app that can show forms asks you in a form.
  • Any other app takes your AI's confirm: true as the yes, so tell your AI to check with you first.
  • An agent with no person to ask needs GUMLOOP_CONFIRM=model, so its confirm: true counts in every app.
  • In a terminal, you add --confirm to the command.

If the answer is no, nothing reaches Gumloop.

Make it read-only

Set GUMLOOP_READ_ONLY=1 and every tool that asks first disappears from the tool list, leaving 50 reading tools.

Keep a log of every write

Set GUMLOOP_AUDIT_LOG to a file path. The server writes one line per attempted write, allowed or blocked, with what it was about to do and who approved it.

Watch out: Agent replies, Brain documents and tool results can hold text from other people, and one can try to give your AI orders. Treat them as data, and use GUMLOOP_READ_ONLY=1 for an agent working on its own.

Your data

Your requests go straight to Gumloop, with no server of ours in between and no tracking. Your key stays in your app's config, your shell or your token file.

Downloads and exports go to new private files, and download links are saved rather than followed. Session files and Brain uploads go to Gumloop, so its own terms apply.

To cut access, delete the key on the Connectors page, and remove the server from your app's config.

Connect more than one Gumloop account

Set GUMLOOP_ACCOUNTS to a JSON list with a name, a key or a token file, and the user and team IDs for each account, and GUMLOOP_DEFAULT_ACCOUNT to the one that acts when a tool names none. Every tool takes an optional account.

An account never borrows another one's key or IDs, so a call for one account can't run as another.

Gumloop MCP server settings

Every setting is an environment variable, in your app's config or your shell.

GUMLOOP_API_KEY
Default
none
What it does
Holds your Gumloop API key
GUMLOOP_TOKEN_FILE
Default
none
What it does
Points at a private file holding the key, instead
GUMLOOP_USER_ID
Default
none
What it does
Holds your user ID, for calls that need one
GUMLOOP_TEAM_ID
Default
none
What it does
Holds your team ID, for team flows
GUMLOOP_ACCOUNTS
Default
none
What it does
Lists several accounts, with their user and team IDs, as JSON
GUMLOOP_DEFAULT_ACCOUNT
Default
the first one
What it does
Picks the account that acts when a tool names none
GUMLOOP_REQUEST_TIMEOUT_MS
Default
30000
What it does
Gives up on a Gumloop request after this many milliseconds
GUMLOOP_MAX_RETRIES
Default
2
What it does
Retries a rate-limited read this many times, and never a write
GUMLOOP_MIN_REQUEST_INTERVAL_MS
Default
150
What it does
Waits this many milliseconds between requests
GUMLOOP_READ_ONLY
Default
off
What it does
1 hides every tool that spends credits or changes something
GUMLOOP_ALLOW_DESTRUCTIVE
Default
on
What it does
0 refuses the tools that ask first
GUMLOOP_AUDIT_LOG
Default
none
What it does
Names a file that gets one line per attempted write
GUMLOOP_SURFACE
Default
full
What it does
search lists three tools that find, describe and run the rest
GUMLOOP_TOOL_TIMEOUT_MS
Default
none
What it does
Gives up on any tool after this many milliseconds
GUMLOOP_CONFIRM
Default
human
What it does
model lets your AI's confirm: true approve on its own, for an agent with no person to ask
GUMLOOP_HTTP_PORT
Default
8787
What it does
Sets the port for --http
GUMLOOP_HTTP_HOST
Default
127.0.0.1
What it does
Sets the address for --http, and any other needs a token
GUMLOOP_HTTP_TOKEN
Default
none
What it does
Sets the bearer token every --http request needs
GUMLOOP_HTTP_ALLOWED_ORIGINS
Default
none
What it does
Lists other websites allowed to call --http, comma separated
GUMLOOP_DEBUG
Default
off
What it does
1 prints debug lines

Troubleshooting

Run the doctor first. It names the step that failed and the fix.

What you seeWhat to do
A 401 or 403 errorThe account isn't on Pro or above, the key was deleted, or your role lacks access. Check the key on the Connectors page
A flow says the user is missingSet GUMLOOP_USER_ID, and GUMLOOP_TEAM_ID for team flows
A flow's inputs are refusedRead the flow's inputs with get_input_schema and use their exact names
A run finished with no outputsOnly a flow's output steps show up as outputs. Add one to the flow in Gumloop
A message gets a 409 errorThe agent is busy. Queue the message, and it goes out after the current turn
A 429 errorYou've hit a limit, like how many agents can run at once. Wait, and check before you try again
"…so it will not run without confirm: true"That's the safety check working. Approve it when your app asks

If the server doesn't show up in your app at all, run the command your app runs, in a terminal, and read the error.

More free tools for automation

Work in Gmail and Google Drive, or run your community in Circle.

Free tools

AI agent readiness checkerSkill generatorClaude prompt generatorPrompt improverDiagram Maker

MCP servers

Google Workspace MCP server & CLIKit MCP server & CLI

FAQs about the Gumloop MCP server & CLI

Here are the questions people ask most about the Gumloop MCP server and CLI.

The Gumloop MCP server & CLI is a free, open source program that lets an AI app like Claude, Codex or Cursor run flows and agents in your Gumloop account.

It has 92 tools for flows, agents, sessions, Brain, skills, MCP servers and evaluations, and the same tools run as a CLI that agents like Claude Code and Codex use, or that you type yourself.

An MCP server is a standard way to give an AI app real access to a tool, so it can act instead of guessing.

MCP stands for Model Context Protocol, and Claude, Codex, Cursor and many other AI apps speak it.

Every server I've made is on my MCP servers page.

The Gumloop CLI is the same program as the Gumloop MCP server, run as commands.

AI agents like Claude Code, Codex and OpenCode run them for you, and you can type them in a terminal or a script.

One npm package installs both.

Use the Gumloop MCP server in an AI app with no terminal, like Claude Desktop's chat, and the CLI in a terminal, a script or a cron job.

In an AI app that can run commands, like Claude Code, Codex or Cursor, the CLI is the lighter choice, because it costs nothing until it runs.

Yes, the Gumloop MCP server is free and open source under the AGPL-3.0-or-later license.

Gumloop's API needs the Pro plan or above, and runs, chats and Brain searches spend Gumloop credits.

You need the Pro plan or above, because Gumloop's API keys come with those plans.

You generate the key on the Connectors page.

A personal key acts as you, with your own connected accounts.

A team key can act as any team member, and each flow step's settings decide whose connected accounts it uses.

Yes, it answers a paused session's approvals, one session at a time, after you approve that answer.

Your AI shows you what the agent wants first.

Yes, the Gumloop MCP server connects as many accounts as you set up in GUMLOOP_ACCOUNTS, each with its own key and IDs.

Every tool takes an optional account, and an account never borrows another one's key.

To disconnect the Gumloop MCP server, delete its key on the Connectors page, which cuts access right away.

Then remove the server from your AI app's config.

The Gumloop MCP server works with any app that speaks MCP.

That includes Claude Code, Claude Desktop, Codex, Cursor, VS Code, Windsurf, Zed, Cline, GitHub Copilot CLI, Gemini CLI, OpenCode, OpenClaw, Antigravity and Hermes.

I did.

I'm Navid Moazzez, and I made the Gumloop MCP server & CLI as one of my free tools on navid.me.

You can read more about me.

The code is on GitHub under the AGPL-3.0-or-later license, and the package is on npm.

More MCP servers & CLIs

View all

Google Workspace MCP server & CLI

Work across Gmail, Drive, Sheets, Docs and Calendar from Claude, Codex or Cursor.

ProductivityGoogle Workspace

Kit MCP server & CLI

Draft newsletters, manage subscribers and tags, and read your email stats from Claude, Codex or Cursor.

NewslettersKit

Circle MCP server & CLI

Manage members and spaces, draft posts, and run events and courses, from Claude, Codex or Cursor.

CommunitiesCircle

Bluesky MCP server & CLI

Post, read and search Bluesky from Claude, Codex or Cursor, or from your terminal.

Social mediablueskyBluesky

Substack MCP server & CLI

Draft posts, publish Notes and check your stats on Substack from Claude, Codex or Cursor.

NewsletterssubstackSubstack

Threads MCP server & CLI

Post, reply and read your insights on Threads from Claude, Codex or Cursor.

Social mediathreadsThreads

Mastodon MCP server & CLI

Post, edit and read your timelines on any Mastodon instance from Claude, Codex or Cursor.

Social mediamastodonMastodon

TikTok MCP server & CLI

Rank your videos, read your stats and publish to TikTok from Claude, Codex or Cursor.

Social mediatiktokTikTok

ThriveCart MCP server & CLI

Check revenue, customers, subscriptions and affiliates in ThriveCart from Claude, Codex or Cursor.

SalesThriveCart

WordPress MCP server & CLI

Edit posts, pages, Elementor and Rank Math SEO on WordPress from Claude, Codex or Cursor.

BlogsWordPress

Midjourney MCP server & CLI

Generate images, follow jobs and download the files from Midjourney with Claude, Codex or Cursor.

AI imagesMidjourney

Google Photos MCP server & CLI

Pick photos, build albums and upload to Google Photos from Claude, Codex or Cursor.

PhotosGoogle Photos

Apple Photos MCP server & CLI

Search, look at and organize your Apple Photos library from Claude, Codex or Cursor, on your Mac.

PhotosApple Photos

Apple Podcasts MCP server & CLI

Track chart ranks, read reviews and search your podcast library from Claude, Codex or Cursor.

Podcastsapple-podcastsApple Podcasts

Podcast Index MCP server & CLI

Search 4 million podcasts and read episode transcripts from Claude, Codex or Cursor.

PodcastsPodcast Index

Google Search Console MCP server & CLI

Read your search traffic, check indexing and submit sitemaps from Claude, Codex or Cursor.

SEOGoogle Search Console

iMessage MCP server & CLI

Catch up, search your history and send iMessages from Claude, Codex or Cursor, on your Mac.

MessagingiMessage

Facebook MCP server & CLI

Read your Page insights, schedule posts and moderate comments, from Claude, Codex or Cursor.

Social mediafacebookFacebook

Facebook Ad Library MCP server & CLI

Research the ads any brand runs on Facebook and Instagram, from Claude, Codex or Cursor.

ResearchfacebookFacebook

Vimeo MCP server & CLI

Read transcripts, file videos into folders and add chapters from Claude, Codex or Cursor.

VideoVimeo

YouTube MCP server & CLI

Read any video's transcript, research channels and manage your own from Claude, Codex or Cursor.

VideoyoutubeYouTube

Telegram MCP server & CLI

Catch up, search years of chats and reply on Telegram from Claude, Codex or Cursor.

MessagingtelegramTelegram

Adobe Firefly MCP server & CLI

Generate and edit images with Image 5, make video and composite products from Claude, Codex or Cursor.

AI imagesAdobe Firefly

OP3 MCP server & CLI

Read your podcast's downloads, real listeners and retention from Claude, Codex or Cursor.

PodcastsOP3

Beehiiv MCP server & CLI

Draft posts, manage subscribers and segments, and read your stats, from Claude, Codex or Cursor.

NewslettersBeehiiv

Calendly MCP server & CLI

Find open times, book and cancel meetings, and read meeting recaps, from Claude, Codex or Cursor.

SalesCalendly

Wistia MCP server & CLI

Upload and organize videos, fix captions and read your analytics, from Claude, Codex or Cursor.

VideoWistia

ScrapeCreators MCP server & CLI

Research public profiles, posts, transcripts and ads across social platforms, from Claude, Codex or Cursor.

Social mediaScrapeCreators

Shotstack MCP server & CLI

Render videos from templates, make AI clips and manage hosted files, from Claude, Codex or Cursor.

VideoShotstack

Buffer MCP server & CLI

Draft and schedule posts, manage your queues and read your analytics, from Claude, Codex or Cursor.

Social mediaBuffer

Cloudflare MCP server & CLI

Manage DNS records, clear the cache and change domain settings, from Claude, Codex or Cursor.

ProductivityCloudflare

Bannerbear MCP server & CLI

Make images, animations and videos from your templates, from Claude, Codex or Cursor.

AI imagesBannerbear

Brandfetch MCP server & CLI

Get any brand's logos, colors and fonts, and compare brands, from Claude, Codex or Cursor.

ResearchBrandfetch

Creatomate MCP server & CLI

Edit templates, check renders for free and make videos and images, from Claude, Codex or Cursor.

VideoCreatomate

Dub MCP server & CLI

Create short links, read clicks and sales, and run your partner program, from Claude, Codex or Cursor.

SalesDub

fal.ai MCP server & CLI

Find models, check prices, and generate images and videos, from Claude, Codex or Cursor.

Videofal.ai

Flodesk MCP server & CLI

Manage subscribers, segments and workflows, and draft campaigns, from Claude, Codex or Cursor.

SalesFlodesk

Fluent WordPress MCP server & CLI

Work with FluentCRM, FluentCommunity and Fluent Forms on your WordPress site, from Claude, Codex or Cursor.

SalesWordPress

Senja MCP server & CLI

Find, approve and tag testimonials, send invites and export them, from Claude, Codex or Cursor.

SalesSenja

Testimonial.to MCP server & CLI

Read, import and export your testimonials, and send request emails, from Claude, Codex or Cursor.

SalesTestimonial.to

Lemon Squeezy MCP server & CLI

Read and refund orders, manage subscriptions and check license keys, from Claude, Codex or Cursor.

SalesLemon Squeezy

Gumroad MCP server & CLI

Read sales and payouts, refund orders and manage products and license keys, from Claude, Codex or Cursor.

SalesGumroad

Fourthwall MCP server & CLI

Manage products and orders, ship them, and run promotions and giveaways, from Claude, Codex or Cursor.

SalesFourthwall

Teachable MCP server & CLI

Read your courses and student progress, and manage enrollments from Claude, Codex or Cursor.

SalesTeachable

Related free tools

All free tools

AI agent readiness checker

See if AI can reach, read and trust your site: 39 checks on 21 pages, the public record, and 15 questions for the AIs.

SEOAny website

Skill generator

Build an Agent Skill for Claude Code, the Claude apps or Codex, with a graded description and test prompts.

AIAny website

Claude prompt generator

Write Claude prompts to Anthropic's own best practices: XML sections, examples in tags, the reason behind each rule.

AIAny website

Prompt improver

Paste a prompt and get it back sharper, to the platform's own rules, with every variable kept.

AIAny website

Diagram Maker

Map workflows and systems with editable diagrams, icons and themes.

Design

X character counter

Counts the way X counts: links as 23, emoji as 2. LinkedIn, Instagram, TikTok, Threads and YouTube limits too.

WritingblueskyBlueskyinstagramInstagram

LinkedIn character counter

LinkedIn's 3,000-character post limit, counted as you type, with X, Threads and Bluesky beside it.

WritinglinkedinLinkedIn

Threads character counter

Threads' 500-character limit, counted as you type, with X, Bluesky and LinkedIn beside it.

WritingthreadsThreads

Bluesky character counter

Bluesky's 300-character limit, counted the way the protocol counts, with X and Threads beside it.

WritingblueskyBluesky

Mastodon character counter

Mastodon's 500-character default, counted the way Mastodon counts links and mentions.

WritingmastodonMastodon

Instagram character counter

Instagram's 2,200-character caption limit, counted as you type, with TikTok and the rest beside it.

WritinginstagramInstagram

TikTok character counter

TikTok's 2,200 caption limit, counted the way TikTok counts it, with emoji as 2.

WritingtiktokTikTok

Navid Moazzez

AI business strategist

Navid Moazzez helps creators and founders master AI and build their own AI Operating System (AI OS) to automate their business and life.

Navid.me is reader-supported. When you buy through links on this site, I may earn an affiliate commission. Read the affiliate disclaimer.

Free AI newsletter

The most actionable AI newsletter for founders

Every week, get proven AI strategies, curated tools, and step-by-step systems to grow your audience, create better content, and build a profitable creator business.

No fluff, no filler, no BS. Just five minutes each week that might level up your online business and life.

P.S. Sign up now to get free access to my ultimate AI tools guide for creators.

Loved by 10,000+ readers
Loved by 10,000+ readers