Connect Re:amaze to ChatGPT via MCP: Support Inbox

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

Re:amaze + ChatGPT

Re:amaze helpdesk from any AI agent: read and answer conversations across email, chat and social channels, post replies and internal notes, manage contacts and their notes, FAQ articles, channels, staff, response templates, satisfaction ratings and support reports.

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:

REAMAZE_BRAND, REAMAZE_LOGIN_EMAIL, REAMAZE_API_TOKEN

Step 1 — Get credentials

Setting up takes about 5 minutes. You need a Re:amaze account with a staff user whose API token the connector uses, and the brand's host name.

1. In Re:amaze open Settings, Developer, API Token and click Generate New Token. Copy the token into REAMAZE_API_TOKEN. Every staff user has their own token, and the connector acts as that user.
2. Set REAMAZE_LOGIN_EMAIL to the email address that same user signs in to Re:amaze with. A token only works together with the email of the user who generated it.
3. Set REAMAZE_BRAND to the brand's host name: the part before .reamaze.io in the brand's address, shown in Settings, Brand Settings. For mybrand.reamaze.io that is mybrand: only the name, made of letters, digits and dashes, without https:// or .reamaze.io.
4. Install the connector. Run reamaze_list_channels to check the connection: it lists the brand's channels with their slugs, which other tools use as category.

Cloud and self-hosted work the same way: the connector signs in with HTTP Basic over HTTPS, the login email as user name and the API token as password. There is no redirect URL and no IP allow list.

One brand per connector: calls go to https://<brand>.reamaze.io/api/v1. Conversations, messages, channels, FAQ articles and response templates are those of that brand; contacts and staff belong to the whole account. Reports and satisfaction ratings cover all brands of the account unless you name one brand. For another brand's inbox, set up a second connector with that brand's host name.

The connector acts as the token's user: what that staff user may see and do in Re:amaze is what the AI may see and do. Replies and notes go out in that user's name unless another sender is given. Reports and satisfaction ratings need the access_reports permission on that user.

Replies go out at once: reamaze_create_message with visibility 0 sends a reply to the customer straight away over the conversation's channel (email, chat, SMS and so on), so the AI shows you the text and waits for your yes. Visibility 1 adds an internal note only staff see. After a staff reply Re:amaze marks the conversation Done unless the AI passes suppress_autoresolve; satisfaction surveys and notifications can be suppressed the same way. A new conversation from reamaze_create_conversation sends the usual notifications unless suppress_notifications is set.

Some updates replace: tag_list in reamaze_update_conversation replaces all tags of the conversation, a new article body replaces the old one, and user_data in reamaze_create_conversation replaces the customer's existing data attributes.

Ids and codes: a conversation is identified by its slug (e.g. knock-knock-a1b2c3) and a channel by its slug (e.g. support); a contact by email address, or by mobile number in E.164 form (+12223334444) with the identifier type mobile; articles by slug, response templates by numeric id. Conversation status: 0 Open, 1 Responded, 2 Done, 3 Spam, 4 Archived, 5 On Hold, 6 Auto-Done, 7 AI Agent Assigned, 8 AI Agent Done, 9 Spam (AI).

Dates, paging and rate limits: list filters take ISO 8601 date-times; reports take two days as YYYY-MM-DD, by default the last 30 days, at most one year apart. Lists return 30 items per page; the AI asks for the next one with page. The API is rate limited per token per minute and answers 429 above the limit: wait and try again.

Common errors: 401 means the email or the token is wrong, often because the email is not that of the user who generated the token. 403 means the token works but its user lacks a permission (reports and ratings need access_reports). 404 means an unknown slug or a wrong brand host name. 422 means Re:amaze refused a create or update, for example a contact that already exists; the answer says why.

What it reads: conversations with subject, status, tags, assignee, channel, custom data and followers, and every message of a thread, replies and internal notes, with sender, recipients and attachments; messages across all conversations by sender, channel, tag or date; contacts with their identities (email, mobile, social accounts) and notes; FAQ articles, published, draft or internal; channels and their settings; staff users and roles; response templates; satisfaction ratings with comments; and reports on volume, response times, staff, tags and channels.

Write tools are on at install: no tool is switched off. The AI can reply to customers and add internal notes (reamaze_create_message), open conversations (reamaze_create_conversation), change their status, tags, assignee or channel (reamaze_update_conversation), create and update contacts and add contact notes (reamaze_create_contact, reamaze_update_contact, reamaze_add_contact_note), and create or edit FAQ articles (reamaze_create_article, reamaze_update_article). Switch off on the connector page the ones you do not want the AI to use.

Not covered: deleting or editing contact notes, editing contact identities, creating staff users, the Status Page API and changes to response templates. No tool deletes anything.

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 ChatGPT

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.

ChatGPT adds MCP servers as plugins on its Plugins page, on the web and on paid plans. Developer mode is no longer needed. In a Business workspace only admins and owners can add one; in Enterprise and Edu an admin can grant the Create plugins with MCPs permission.

1. Go to chatgpt.com/plugins, click Add, then Add custom MCP server.
2. Enter a name, e.g. Name: Re:amaze.
3. Under Connection, enter the Server URL https://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID and keep Authentication on OAuth.
4. Select I understand and want to continue, then Create as a plugin. When ChatGPT asks, sign in to AnythingMCP and allow access.
5. Install the new plugin from your personal plugins, then type @ in a chat and pick it.

Available tools

Example prompts

• "Which open conversations are unassigned, and what does each customer want?"
• "Show me the latest thread with [email protected] and suggest a reply based on our FAQ articles; send it only after I approve."
• "Add an internal note to the late-delivery conversation saying we have contacted the carrier, and put it on hold until Monday."
• "How did our response times develop over the last 30 days, and who on the team sent the most responses?"
• "List last month's satisfaction ratings of 1 or 2 stars with their comments."

FAQ

Does the free ChatGPT plan work with custom MCP servers? OpenAI documents custom MCP servers for paid plans only; the Free plan is not on its list.

Next steps

• Connector landing page
• Connect to Claude