MCP · Anthropic
Marketing MCP for Claude
Add Spreeflo to Claude as a custom connector and it can build segments, draft campaigns, and pull performance — in a quick question or a longer task that keeps running in the cloud.
Web, desktop and mobile·Connects from Anthropic’s cloud·Every Claude plan, Free included
https://mcp.spreeflo.com/mcpWrites apply to your live workspace immediately — there's no approval queue here. Keep a human in the loop and check Claude's confirmation settings first.
The agent
What is Claude?
Claude is Anthropic’s assistant. Since September 2026, what used to be Claude Cowork — multi-step tasks across your files and connected apps — is part of every Claude conversation rather than a separate mode. Custom connectors let Claude reach remote MCP servers, which is how it reaches Spreeflo.
Cloud-side connection
Claude dials your MCP server from Anthropic’s infrastructure, not your laptop — so the connection keeps working when your machine is asleep.
Multi-step autonomy
Any conversation can become a task. Claude chains tool calls on its own, so “segment these people then draft them an email” is one request rather than four.
Alongside your other tools
Spreeflo sits next to whatever else you have connected, so your marketing data can meet the rest of your work in one task.
Setup
Connect Claude in about a minute
One entry in Claude, then sign in to Spreeflo. No key to create or paste.
Add the Spreeflo MCP server
Custom connector- 1.Open Customize → Connectors, click Add and choose Add custom connector.
- 2.Name it Spreeflo, paste the URL below, and click Add. Leave any client ID and secret fields empty.
- 3.Click Connect on the new connector to start the Spreeflo sign-in.
https://mcp.spreeflo.com/mcpOn Team and Enterprise, an Owner adds the connector under Organization settings → Connectors, and members then click Connect under Customize → Connectors.
Sign in and approve
Spreeflo opens in your browser. Sign in, pick the workspace Claude should reach, and click Approve. Tick Make changes if it should be able to write — leave it off for a read-only connection.
Check it works
Ask Claude for something small — "list my Spreeflo audiences". A working connection is one where the connector returns your real audience names rather than an auth error.
Full setup guide and troubleshootingIn practice
What Claude does with it
Two ordinary requests, and the Spreeflo tools behind them.
Pull last week’s campaign numbers and tell me which segment is going cold.
Compared performance across the window, then previewed the segments behind the weakest results:
list_campaignsget_campaign_performancelist_segmentsget_segment_preview
All reads — nothing in the workspace changed.
Take the positioning doc I shared and draft a re-engagement email in our voice.
Read the brief, checked your business profile for tone, and saved a draft template:
get_business_profileread_format_doccreate_email_template
Saved as a draft. Publishing and sending stay in the Spreeflo app.
Running it
What changes when the connection is cloud-side
Claude does not proxy through your laptop — Anthropic’s infrastructure holds the connector and calls Spreeflo directly. That is what makes it convenient, and it is also what makes these four things worth knowing.
The sign-in is stored with the connector
Because the call originates in Anthropic’s cloud rather than on your machine, the connection you approved lives with the connector configuration. It is scoped to the one workspace you picked and acts as you — so anyone with access to your Claude account reaches that workspace too.
Read-only is a choice at approval
The approval screen asks whether Claude may make changes. Leave it unticked and Claude is offered only the tools that read — a sensible default for a connector that mostly reports. To upgrade later, reconnect and tick it.
One request can become many writes
Claude chains tool calls, so “clean up the segments and redraft the welcome email” may touch a dozen objects before it reports. Two settings slow that down: the connector’s Tool permissions, where Spreeflo’s write tools can be set to Needs approval, and the conversation’s Manual mode, where Claude asks before each action. Spreeflo applies each call as it arrives.
When the connector will not add
A connector that fails at the point of adding is nearly always a mistyped URL — a missing /mcp or an extra path segment. Re-paste the endpoint exactly as shown above. If the Spreeflo page says the link has expired, start the connection again from Claude: an approval request lasts 15 minutes.
It keeps working when you are not there
Longer tasks run in the cloud, so they carry on after you close the laptop or leave the page — genuinely useful and easy to forget. A task you kicked off and walked away from keeps running against your live workspace. If you want a hard stop, disconnecting Spreeflo under Customize → Connectors is the switch that works — closing the app is not.
Why this pairing
Claude + Spreeflo
- Connects without installing anything locally
- Reaches Spreeflo alongside your files and other connectors
- Chains multiple Spreeflo tools in one request
- Signs in with your Spreeflo account — no key to paste
- No tool for irreversible actions, not left to settings
Coverage
Everything it can touch
The same surfaces you work with in the app, exposed as tools — and a short list that is deliberately off limits.
Claude usually spans the widest slice of this in a single task, since it chains calls — reading performance, then acting on what it found, without stopping to ask in between.
Audience
- Audiences and contacts
- Segments, static and dynamic
- Tags
- Custom contact attributes
- CSV import
Sending
- Campaigns and journeys
- Email templates
- Web push templates
- Sender and domain identities
Capture
- Forms and question types
- Inline and popup embeds
- Response summaries
- CSV export of responses
Insight
- Campaign and web push performance
- Tracked events and pages
- Link clicks
- Form analytics
What MCP can’t do
These are irreversible or they reach real people, so they are removed from the surface rather than left to your client's confirmation settings. A person does them in the app.
Supercharge your marketing automation
FAQ
Questions, answered.
In September 2026 Anthropic merged Cowork and chat into one Claude: there is no separate Cowork mode to pick any more, and everything Cowork did is available from any conversation. Connectors carried over, so a Spreeflo connector added in Cowork keeps working. Enterprise organisations may still see Cowork listed separately while the unified experience is in beta there. This page used to live at /claude-cowork.
No. Claude calls Spreeflo from Anthropic’s infrastructure, and longer tasks run in the cloud, so a Spreeflo task keeps going with your laptop shut. Claude Desktop only has to stay open for work that touches your local files, its browser, or computer use — none of which a Spreeflo call needs.
Yes. Every Claude plan can add custom connectors, and Free accounts can add one — so Spreeflo can be that one. Pro, Max, Team and Enterprise have no such cap. Nothing on the Spreeflo side changes with the Claude plan: the same tools and the same rate limit apply.
No. Spreeflo runs a hosted MCP server, so Claude calls it directly over Streamable HTTP — nothing to install, nothing to self-host, no SDK to keep up to date. A server URL and a sign-in (or, for a headless agent, an API key in the Authorization header) are the whole integration.
It cannot delete anything or start a send. Spreeflo has no MCP tool for deletes, starting or resuming a campaign or journey, sending or scheduling a broadcast, or publishing a form, rather than leaving them to client settings. Those stay in the Spreeflo app, where a person does them. Journeys you have already switched on keep running, though, so editing a contact, its tags, a segment or an email template, or importing contacts, can set one off, just as the same edit would in the app.
The key resolves to exactly one workspace, and every action runs as the person who owns that key — their membership re-verified on each request. The audit trail names a real person, so agent-made changes are attributable.