Connect Marqeable to ChatGPT: Setup, Permissions, First Test
Short answer (published October 22, 2026): Marqeable’s connector is an MCP server at https://app.marqeable.com/mcp. In ChatGPT, turn on Developer mode, create a connector with that address and OAuth, and sign in with your normal Marqeable login. Set its permission to Allow read-only tools. ChatGPT can then read your calendar, campaigns, content and automations and write drafts into Marqeable. It cannot send, publish, launch or delete anything, under any setting.
ChatGPT with a Custom GPT knows what you told it about your marketing the day you wrote the instructions. ChatGPT with a connector knows what is on your calendar this morning. This post is the setup for the second kind: how to add Marqeable’s connector, which permission level to choose, which tools to let run and which to hold, and a first test with the result you should expect. It is written from our own runs in October 2026; OpenAI moves these settings between releases, so where the path differs, search Settings for “Developer mode”.
Marqeable is an independent company. ChatGPT is OpenAI’s product, and Marqeable is not made by, endorsed by, or affiliated with OpenAI. What we provide is a connector their assistant can use.
What you need
| Requirement | Detail, as of October 2026 |
|---|---|
| A ChatGPT plan with Developer mode | Custom MCP connectors need Developer mode, which is on paid plans. In a Business workspace an admin may need to allowlist the connector first |
| A Marqeable account | You sign in as yourself. ChatGPT sees exactly what your login can see, on the account you have selected |
| A browser or the desktop app | Setup is in Settings; it is not available in every mobile build |
Step 1: Add the connector
The server address is:
https://app.marqeable.com/mcpIt is a remote server over HTTP with OAuth 2.1. No API key is created, copied or stored anywhere.
- In ChatGPT, open Settings and turn on Developer mode. OpenAI’s Developer mode guide has the current location; it has lived under Connectors, Apps and Plugins at different points this year.
- Create a new connector. Name it Marqeable and paste the address.
- Choose OAuth as the sign-in method and save.
- Sign in to Marqeable in the browser window that opens. If you belong to more than one Marqeable account, the connector uses the one you last selected in the app.
Step 2: Confirm it works
Open a normal chat, enable the Marqeable connector from the tools menu, and ask:
Run whoami and tell me which account I am signed in to.
The reply names the signed-in email and the current Marqeable account. If it does not, see troubleshooting below before going further. Nothing else in this post works until this does.
Step 3: Choose the permission level
Under Settings, each connector has one permission level (recent releases list connectors under Plugins; earlier ones under Connectors). ChatGPT’s own wording:
| Level | What it means |
|---|---|
| Always ask | Ask before reading or making changes |
| Allow read-only tools | Read without asking, but ask before making changes |
| Allow low-risk tools (default) | Automatically approve low-risk tools, but sensitive tools may be denied |
| Allow all tools (marked “elevated risk”) | Use tools without asking |
Set Marqeable to Allow read-only tools. Marqeable’s server marks every tool as read-only or not, and this level reads that flag: reads run freely, anything that creates or changes something asks first.
One result from our own testing that you should know. With that level set, we asked ChatGPT in the conversation to create one campaign draft, and it did so without a separate prompt. An explicit request in the chat counts as the approval. So the permission level protects you from the assistant acting on its own initiative; it does not second-guess what you ask for directly. That is the right way round, and it is why the next section matters more than the setting.
Step 4: Know which tools do what
The connector’s tools fall into four groups. Reads can run all day. Drafts are safe by design. A few tools make direct changes. Two should always be yours.
Reads: let them run
These change nothing. They are what makes a connector better than a Custom GPT: the answer reflects the account right now.
| Tool | What it reads |
|---|---|
whoami, list_accounts | Who is signed in, and to which account |
get_business_context | Current focus, offers, audience and voice |
get_content_calendar, get_marketing_summary | Calendar entries, planned content, suggested dates, the current plan |
list_campaigns, get_campaign, get_campaign_results | Campaigns, their pieces, their numbers |
search_content, get_content, list_content_blocks | Content pieces, briefs, current copy, saved blocks |
list_content_templates, search_image_library, get_image_usages | Templates, existing images, where each image is used |
get_automation_overview, get_automation_plan, get_automation_schedule | Automations, their steps, their timing |
get_event_catalog, inspect_landing_pages, get_landing_page, get_landing_page_analytics, list_form_submissions | What an automation can start on, page states, page numbers, form entries |
check_account_setup, check_campaign_readiness, check_automation_readiness, validate_automation_plan, preview_automation, preview_audience_reach, get_sending_limits, get_open_issues | Checklists, dry runs and anything that is currently blocking a send |
get_brand_look | The account’s email look |
Drafts and suggestions: safe to allow
Each of these creates or edits something a person reviews in Marqeable. Nothing reaches a customer.
| Tool | What it creates |
|---|---|
create_campaign_draft, plan_campaign_pieces, add_planned_pieces, remove_campaign_piece, update_campaign_brief, set_campaign_audience, set_campaign_tags | A campaign in draft and the shape of its pieces |
create_content_draft, start_content_build, stop_content_build | A content piece in draft, and its build |
create_automation_draft, update_automation_draft, create_automation_step_content | An automation draft, a proposed revision, draft content for its empty steps |
add_content_suggestion, request_content_changes, insert_content_block | A suggested edit, a change request at a checkpoint, a saved block placed into a draft |
update_content_brief, answer_open_items | Changes to a draft’s brief, answers to the questions a build raised |
update_landing_page, send_test_submission | Edits to a Draft landing page, and a test of its form that creates nothing |
A content build runs for about a minute and can generate images, so a sensible habit is to ask ChatGPT what images it plans before it starts one.
Direct changes: ask first
These take effect without a review step in Marqeable, so the review has to happen in the conversation.
| Tool | Why it needs a yes |
|---|---|
add_calendar_entry, update_calendar_entry, adopt_suggested_date, dismiss_suggested_date | Calendar changes feed the next plan directly |
save_content_as_template, set_default_content_template, switch_content_template | Changes what future pieces start from |
update_brand_look, reset_brand_look | Changes the look of every email |
save_content_block | Adds to the blocks everyone can insert |
unpublish_landing_page | Takes a live page down |
decide_automation_matching_hold | Decides what happens to a held automation run |
retry_automation_post | The one tool that acts immediately: it posts a failed or held-back automation post. Confirm every time |
Yours alone
| Tool | Why |
|---|---|
set_feature_enabled | Turns account features on or off. A person’s decision |
switch_account | Only when you explicitly ask for it, and check whoami after |
Never sign in to Marqeable inside a browser that ChatGPT controls, and never paste a password into a chat. If the assistant proposes “the manual route” because a tool does not do exactly what you asked, tell it to use only the connector and rephrase the request in terms of what the tools can do.
What the connector cannot do, under any setting
There is no tool to send a message, publish content, launch a campaign or turn on an automation. There is no delete tool. A permission level cannot allow an action that has no tool, and that is the point: settings drift, the tool list does not. The longer argument is in the draft-only rule.
The honest caveat is the direct-changes table above. Calendar, template, brand-look and landing-page tools make real changes, and one tool retries a post. None of them reaches a new customer, but they are why the ask-first group exists.
A first test
Give it a small job with a visible result:
You: Check my marketing calendar for the next four weeks against my current focus in Marqeable. If there is a gap, draft one campaign to fill it and give me the link. Do not change the calendar.
Expected result: a handful of reads (
whoami,get_business_context,get_content_calendar,list_campaigns,get_automation_overview), one campaign draft, and a link to review it in Marqeable. The draft sits in Draft status with nothing built, sent or scheduled.
Then ask: “List the exact Marqeable tool calls you made, in order.” The answer, down to the arguments, is your audit of what happened.
Troubleshooting
| What you see | What to try |
|---|---|
| No Developer mode switch | Check your plan, or ask your workspace admin to enable it or allowlist the connector |
| Sign-in succeeds but no tools appear | Remove and re-add the connector, then start a new chat |
| ChatGPT asks before every read | Set the connector’s permission to Allow read-only tools |
| ChatGPT wants to open Marqeable’s website | Tell it to use only the connector. Never hand over a password |
| Wrong account | Ask it to run whoami. Switch accounts yourself, deliberately, and run whoami again |
| A write “cannot be confirmed” | Some plans confirm write actions per conversation; approve it once and retry |
The same server works from Claude (setup guide), from Claude Code and Codex in a terminal (Codex MCP for marketing), and from an always-on dot (connecting Marqeable to OpenAI dots). The connection is to the account, not to the client. If you are choosing between a Custom GPT and a connector, that comparison is the place to start.
Frequently asked questions
How do I connect Marqeable to ChatGPT?
Turn on Developer mode in ChatGPT Settings, create a connector named Marqeable with the address https://app.marqeable.com/mcp, choose OAuth, save, and sign in with your normal Marqeable login. Then open a chat, enable the connector and ask it to run whoami.
What can ChatGPT do in Marqeable once connected?
Read your business context, campaigns, content, calendar, automations, templates, image library and landing pages, and create drafts: campaigns, content, automations, brief edits, suggestions and change requests. A few tools change the calendar, brand look or templates directly. There is no tool to send, publish, launch or delete anything.
Which permission level should I use?
Allow read-only tools. Reads run without asking; anything that creates or changes something asks first. An explicit request in the conversation counts as approval, so the setting guards against the assistant acting on its own initiative, not against what you ask for directly.
Can ChatGPT send my campaigns through Marqeable?
No. There is no send, publish, launch or delete tool, and no setting can add one. ChatGPT drafts; a person reviews and launches in Marqeable.
The bottom line
Four steps: Developer mode, a connector with the address and OAuth, your own login, and the read-only permission level. After that ChatGPT works from the account as it is today, not from a snapshot you wrote in August, and everything it writes lands as a draft with a link to the page where you approve it. The settings protect you from initiative; the tool list protects you from everything else.
See where the drafts go: Marqeable’s agents build and run your campaigns, run behavior-triggered automations across text and email, answer and qualify every visitor with AI website chat, and tie revenue to the exact message with attribution.
Marqeable runs your campaigns, answers every visitor, text, and email in seconds, and turns them into booked jobs and meetings - even at 9pm on a Saturday. We’re in private beta with a small early cohort. Get early access
