MCP product preview
Connect your AI assistant to Moosewave with clear limits.
Model Context Protocol, or MCP, is a standard connection that lets a compatible AI assistant use approved actions from another product. This preview shows how an assistant could work in Moosewave while staying within the access you give it.
The examples cover campaign inspection, drafting, sender and inbox checks, migration previews, and result review. They show the intended workspace, permission, input, and approval rules.
You ask
“What should we fix before tomorrow’s campaign?”
- 01ReadCampaign details
- 02CheckInbox placement
- 03ReturnAnswer with its sources
Your AI client, Moosewave tools
Give the agent a job, not unlimited access.
Type a job below to open it in the guided workspace. Each connection can use only the parts of Moosewave you allow, while approvals and activity stay beside the work.
- 01InspectAllowed information only
- 02PlanActions and limits
- 03ApproveExact proposed change
- 04RecordRequests and results
A connection, not another chatbot
Keep your preferred assistant. Give it a clearer way to work.
Normally, an assistant knows only what you paste into a chat. MCP can give it specific product actions: read this campaign, prepare that draft, check this audience, or request that action. In the model shown here, the result comes from the product rather than from guesswork.
Moosewave remains in charge of product access: it decides what the connection may see, checks each request, applies workspace rules, and records what happened.
- 01
Ask in plain language
Begin with the outcome, not a developer guide.
- 02
Review work in Moosewave
The design returns drafts and the information used to the shared workspace.
- 03
Follow every request
The planned history keeps the action, connection, timing, and result visible.
Work you can review before it runs
Start with a job your team already does.
Choose an illustrative example, then move through it one step at a time. It shows the assistant making a request, Moosewave checking it, the workspace saving what happened, and the team keeping the important decision.
Read-only investigation
Turn a vague result into a useful next step.
This preview shows an assistant opening a campaign, reviewing where test emails landed, checking the sending setup, and organising what a team could change.
The example investigation reads records but changes nothing.
Why did our last campaign struggle?
Read the campaign, audience, and delivery record
Moosewave returns the selected campaign and its recorded result, so the answer starts from the right send rather than a pasted screenshot.
The assistant is looking at the requested campaign
- Open the campaignWorking now
- Check inbox placementWaiting
- Rank the next fixesWaiting
Three practical MCP examplesOpen the campaign, migration, and reporting examples when you want to see each request, check, and result.
Three places to begin
The question is simple. The work behind it is real.
Each example begins with a normal request, shows the steps Moosewave would take, and says what still needs a person.
- 01
Read-only investigation
Diagnose a campaign
“Why did our last campaign struggle?”
This preview shows an assistant opening a campaign, reviewing where test emails landed, checking the sending setup, and organising what a team could change.
- 01Open the campaign
- 02Check inbox placement
- 03Rank the next fixes
What stays protectedThe example investigation reads records but changes nothing.
Why it helpsA diagnosis with links to the information used to reach it.
- 02
Draft before impact
Prepare a journey
“Prepare a welcome journey in our voice and show me every message before anything is scheduled.”
This preview shows an assistant using your saved voice, templates, and audience rules to prepare a welcome series for team review inside Moosewave.
- 01Read the brief
- 02Prepare the drafts
- 03Review in Moosewave
What stays protectedThe example keeps creating a draft and sending it as separate actions. Team approval rules still apply to sends.
Why it helpsThe drafts arrive in Moosewave for the team to review and continue.
- 03
Preview before moving
Preview a migration
“Show me what would move from our current platform before we commit.”
This preview shows an assistant listing what is in an old account, choosing which records to include, previewing the move, and explaining anything that would be left out before a person starts it.
- 01Review the old account
- 02Preview without importing
- 03Review before the move
What stays protectedLooking through the old account and previewing the move do not change the new account. Starting the move is a separate action with stricter permission.
Why it helpsYour team can inspect the move before any records are imported.
MCP setup and control detailsOpen the proposed tool areas, allowed actions, connection choices, and what your AI provider may store when you need technical depth.
Previewed tool areas
The tool catalog follows the work, not the company structure.
The preview shows where MCP could help across a connected marketing workflow. It is not a list of tools available today: actual access would depend on the app, connection method, permissions, whether changes are allowed, and the product feature.
Campaigns & templates
Find and adapt templates, prepare drafts, run final send checks, compare campaigns, and inspect the result.
Audiences & consent
Find subscriber details, preview audiences, work with lists and tags, and make the do-not-contact list take priority over targeting.
Deliverability
Check sender authentication, reputation, links, rendering, and test-mailbox results before deciding what to change.
Journeys & channels
Prepare automation flows, email, SMS, forms, and send-time decisions with the right channel rules in force.
Analytics & integrations
Read campaign results, compare source records, check whether connections are working, and bring updates from other tools into the same workflow.
Migration
List what is in the old account, preview which records would move, review the checks, watch progress, and keep a way back if something goes wrong.
Go deeper into how Moosewave handles deliverability, analytics, automation, and audience control, or follow the connected ecosystem from idea to result.
Useful access, clear limits
MCP is not a shortcut around your workspace.
The design treats assistant access like access to a live workspace: keep it limited, check it when an action runs, and show what happened afterward. This page explains that model; it does not provide a live connection.
- 1Access
One access key, one small job
Each connection has its own access key, starts read-only, and receives only the permissions the job needs.
- 2Checks
The same product rules
Moosewave checks each request against the same workspace permission rules and do-not-contact list, so an AI app cannot quietly act as an administrator.
- 3Decision
Review before a real change
Drafting and sending stay separate. Teams can preview changes and require approval before data changes or a message goes out.
- 4History
A history you can follow
The design shown here would record the action, connection, result, and time, with important changes visible in the product activity log.
From question to recorded result
The intended connection has five clear steps.
The planned setup starts with a compatible app and limited workspace access, then separates what it may read or change from the important actions a team must review.
- 01
Choose a compatible app.
The preview covers desktop assistants, editors, remote clients, and private team tools. Confirm which apps and connection methods are currently supported.
- 02
Choose what it may access.
Each connection gets its own access key and only the permissions it needs. Permission to change data can remain off.
- 03
Ask for an outcome.
The app requests a specific Moosewave action. Moosewave checks the request and the connection’s permission before continuing.
- 04
Review important changes.
Reading, previewing a change, changing data, and sending are separate actions. Your approval rules still apply.
- 05
Follow the history.
The activity history shows what was requested, what happened, which connection was used, when it happened, and the result. You can remove the connection’s access.
Connection on your computer
For compatible desktop assistants and editors, with the connection running alongside the app.
Private web connection
For compatible apps that can sign in to a protected MCP service over the web.
Private team tools
For shared assistants and internal tools inside a team environment protected by sign-in.
What your AI provider sees
In the intended model, Moosewave controls workspace access. Your chosen AI app controls what happens on its side.
A deployed connection would let Moosewave decide which actions the app may request and which results it may receive. The app or AI provider decides how prompts and returned information are handled, stored, or used. Review those privacy and storage settings before making a real connection.
Review Moosewave’s current security and data-protection posture while evaluating the preview. The public site does not provide sign-in details for a live workspace.
To understand who may read, draft, recommend, schedule, or send, read the agentic email permissions blueprint.
When your product, not an assistant, needs to send a predictable message such as a password reset, receipt, or security alert, use the Moosewave transactional email API. The MCP model shown here is the proposed entry point for limited assistant workflows.
For an ongoing connection that moves provider records or events, browse the Moosewave integrations directory. This MCP preview shows how an assistant could request an action; an integration maintains the data connection itself.
Questions teams ask about the MCP preview
The public site does not provide sign-in details or open an MCP connection. The answers below explain the intended setup and the checks to make before implementation.
What is Model Context Protocol (MCP)?
MCP is a standard way for an AI app, such as an assistant or editor, to ask another product to do specific jobs. This Moosewave preview shows proposed actions for campaigns, audiences, email delivery, automations, analytics, integrations, and migrations. It is not a list of tools available today.
Do I need to be a developer to use Moosewave MCP?
The intended everyday experience does not require development work after setup: a marketer asks in plain language and reviews the result in Moosewave. An administrator or developer still needs to make the first connection. The public site does not currently provide sign-in details or setup.
Which AI clients can connect to Moosewave MCP?
The preview covers desktop assistants, code editors, remote apps, and private team tools. It does not promise that a particular app or connection method is available today. Check current compatibility with Moosewave and the MCP app you choose.
Can an MCP client send a campaign or change customer data?
Not from this public preview: it supplies neither a live connection nor sign-in details. The intended product model prevents data changes unless they are explicitly allowed, separates drafts from sends, and applies the workspace’s approval rules.
Does MCP bypass Moosewave approvals?
The intended model does not bypass Moosewave approvals. An action covered by an approval rule follows that path. Because not every data change requires approval, an administrator should choose permissions and workspace rules carefully.
Where does my data go when I use an AI client?
In the intended model, Moosewave returns only information allowed for that connection and keeps a history of the request and result. The AI app controls how prompts and returned information are handled, stored, or used. No customer data is exchanged through the public page itself.
Can our team run the MCP connection privately?
The intended setup includes connections on your own computer and a private web connection protected by sign-in. The public preview does not provide sign-in details or a live connection; check current app and setup availability before planning use.
Does MCP replace the Moosewave API or other developer tools?
No. MCP is a distinct way for an AI app to request product actions. APIs and SDKs help developers build product connections; command-line tools (CLIs) run tasks from a terminal; webhooks send automatic event updates. This preview does not claim that any of those options are available today.
See how an assistant could work with Moosewave.
Use the guided product preview to understand the workflow. MCP sign-in details are not provided here; check current connection and action availability before implementation. Get product and signup updates; joining the email list does not request product access.