Claude + DigitalOcean: What the Integration Can (and Can't) Do in 2026
Yes — DigitalOcean ships an official MCP server you can add to Claude. Two caveats up front: DigitalOcean is not in Anthropic’s connectors directory (claude.com/connectors/digitalocean returns a 404), and its remote MCP server authenticates with a DigitalOcean API token in the header — not OAuth — so it fits Claude Code or any client that supports custom headers, but it can’t do the browser-OAuth flow that claude.ai web custom connectors expect. Wire it up and Claude can manage droplets, App Platform, databases, and Kubernetes from a chat. The catch is the usual one: it only works inside a conversation you start. There are no triggers, nothing watches your infra, and nothing happens while you’re away.
Here’s exactly what the integration does, how to turn it on, where the limits bite, and what to use if you want DigitalOcean work that runs on its own.
What the DigitalOcean MCP server does
The server connects Claude to your DigitalOcean account across nine services — Accounts, App Platform, Databases, DOKS (Kubernetes), Droplets, Insights, Marketplace, Networking, and Spaces — exposed as tools you drive in natural language.
In practice, Claude can:
- Manage droplets — create, resize, and inspect droplets.
- Deploy on App Platform — create and update apps and read deploy status.
- Work with managed databases and Kubernetes — provision and inspect managed DBs and DOKS clusters.
- Handle networking and Spaces — manage networking resources and object storage.
Read/write scope follows your token’s scopes — DigitalOcean supports granular, custom-scoped tokens, so “read-only” just means issuing a read-only token.
How to set it up
The maintained multi-service server is @digitalocean/mcp, and DigitalOcean publishes its own Claude Code tutorial:
- Create a DigitalOcean personal access token (pick the scopes you want — granular scopes are supported).
- Add it in Claude Code:
claude mcp add digitalocean-mcp -e DIGITALOCEAN_API_TOKEN=... -- npx @digitalocean/mcp(the official repo is underdigitalocean-labs; docs at docs.digitalocean.com/reference/mcp/). - For the remote server, use a per-service endpoint — e.g.
https://apps.mcp.digitalocean.com/mcpfor App Platform — with the token passed via an Authorization header (--header). Because it’s token-auth, not OAuth, this works in Claude Code and header-capable clients, not the claude.ai web “add custom connector” OAuth flow.
Two naming gotchas: the multi-service repo is digitalocean-labs/mcp-digitalocean — the older digitalocean/digitalocean-mcp is App Platform-only. And the AI platform is now the Gradient AI Platform (“GenAI Platform” is stale branding).
The limits that actually matter
The server is genuinely good at pulling DigitalOcean into a conversation. But its shape is “an operator you direct,” not “an agent that runs.” Three limits define it:
- No triggers, no monitoring. The server only works inside a conversation you start. There’s no “restart the droplet if CPU pins” or “alert me when a deploy fails.” And DigitalOcean has no general webhook system for account resources — monitoring alerts go to email or Slack, and polling the API is the automation pattern. Nothing fires into the MCP server — you have to be there, prompting.
- Conversation-only. Claude manages a resource when you ask; it doesn’t sit watching your account and reacting on its own. Close the chat and nothing continues.
- Laptop-bound for anything scheduled. The closest thing to “running on its own” is Claude Cowork’s scheduled tasks, which fire on a fixed clock, not on inbox events. That’s not an always-on infra agent.
So Claude is great for “create this droplet and deploy the app” — not for “watch the fleet and act when a metric crosses a line.”
If you want DigitalOcean work that runs on its own: Carly
The moment you want something to happen around DigitalOcean without you in the chat — check a service on a schedule, email the team when a droplet looks unhealthy, kick off a routine and route the result — you’ve crossed past what Claude’s MCP server is for.
That’s where Carly fits. Carly is an AI executive assistant built to run on a schedule and act, not just answer in a chat:
- Runs 24/7 in the cloud — since DigitalOcean has no webhooks, Carly polls the API on a schedule (health checks, deploy status, resource counts) and acts on what it finds; your laptop doesn’t need to be awake.
- Connects DigitalOcean to the rest of your stack — take a scheduled check and turn it into an email, a task, a CRM note, or a calendar block in the same workflow.
- Actually sends and updates — drafts and sends email (Gmail and Outlook) with attachments, files and labels, manages tasks, updates your CRM, and records meetings.
- Builds the workflow for you — tell it “every morning, check my droplets and App Platform status and email me a summary” in plain English; it interviews you, then builds it with you. No prompt engineering.
AI agents start at $35/month, and steps in a workflow that don’t use AI run free and unlimited. Carly connects to 200+ tools across 40+ categories — see integrations. DigitalOcean connects to Carly via your own API key — paste your DigitalOcean personal access token on carlyassistant.com/integrations, and Carly can do whatever the DigitalOcean API allows.
Claude vs Carly
| Claude (DigitalOcean MCP) | Carly | |
|---|---|---|
| Manage droplets & apps | Yes | Yes |
| Work with managed DBs & Kubernetes | Yes | Yes |
| Runs on a schedule | No | Yes |
| Polls & acts on infra status on its own | No | Yes |
| Works while laptop is closed | No | Yes (cloud) |
| Sends a summary as email | No (Gmail draft-only) | Yes (Gmail + Outlook) |
| Pricing | Pro $20 / Max $100–$200 | AI agents from $35/mo |
Claude’s MCP server is a strong DigitalOcean control panel inside a chat. Carly is a teammate that checks your infra on a schedule and acts on it.
Frequently Asked Questions
Does Claude work with DigitalOcean?
Yes, through DigitalOcean’s official MCP server — but it’s a vendor MCP, not an entry in Anthropic’s directory (claude.com/connectors/digitalocean is a 404). The remote server uses API-token header auth, not OAuth, so you add it in Claude Code or a header-capable client (not the claude.ai web OAuth connector flow). Once wired up, Claude can manage droplets, App Platform, databases, and Kubernetes inside a chat.
How do I connect Claude to DigitalOcean?
Create a DigitalOcean personal access token, then add the server in Claude Code: claude mcp add digitalocean-mcp -e DIGITALOCEAN_API_TOKEN=... -- npx @digitalocean/mcp. For the remote server, use a per-service endpoint like https://apps.mcp.digitalocean.com/mcp and pass the token in an Authorization header. The maintained repo is digitalocean-labs/mcp-digitalocean.
Can Claude react to DigitalOcean events automatically?
No. The MCP server works inside a conversation you start, and DigitalOcean has no general webhook system for account resources — polling the API is the pattern. For scheduled checks that act on their own, you need an agent platform like Carly, which runs 24/7 in the cloud.
Can I add the remote DigitalOcean MCP to claude.ai as a custom connector?
Not the usual way — the claude.ai web custom-connector flow expects OAuth, and DigitalOcean’s remote MCP authenticates with an API token in the header. Use Claude Code (or another client that supports custom headers) instead, or run the local @digitalocean/mcp server.
More: Claude connectors · Claude + Supabase · Can Claude send emails · Claude vs Carly · Claude Cowork alternatives · Best AI tools for solopreneurs
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
