Split MCP server

What the model can do

Split (Harness FME) feature flags: read projects, environments, flags, what each flag serves in an environment and segments with their size, and kill or restore a flag, create, describe, tag and delete flags, submit and approve, reject or withdraw change requests, add keys to a segment and create segments.

Reads

• split_list_workspacesSplit projects (called workspaces in the API) with their ids, which every other Split tool needs.
• split_list_environmentsThe environments (Production, Staging…) of a Split project, with ids.
• split_list_flagsFeature flags in a Split project: name, description, traffic type, tags, rollout status and when created.
• split_flag_definitionsWhat Split flags serve in one environment: whether killed, treatment names, default treatment, traffic allocation, how many targeting rules…
• split_list_segmentsSegments in a Split project (or those active in one environment): name, description, traffic type, tags, when created.
• split_segment_sizeHow many keys (users, accounts…) a Split segment holds in one environment.

Acts, when you ask

These change something in Split. The model is told to use them only when you ask for exactly that, and each one is granted by name: per team on a shared connection, by the owner on a personal one.

• split_kill_flagKills a Split feature flag in one environment: every SDK immediately serves the flag's default treatment to everyone, overriding its targeting.
• split_restore_flagRestores a killed Split feature flag in one environment, so its targeting rules apply again for live users.
• split_create_flagCreates a Split feature flag for a traffic type (such as user or account).
• split_update_flagChanges a Split flag's description, adds a tag to it, or sets its rollout status (by id).
• split_delete_flagPermanently deletes a Split feature flag: its definition is removed from every environment and any SDK still asking for it gets "control".
• split_submit_change_requestSubmits a Split change request to kill or restore a flag in one environment, for an environment where changes need approval.
• split_update_change_request_statusApproves or rejects a Split change request this key was named an approver on, or withdraws one this key submitted.
• split_add_segment_keysAdds keys (user or account ids) to an existing Split segment in one environment; flags targeting the segment serve them accordingly at once.
• split_create_segmentCreates an empty Split segment for a traffic type.

Connect it

First connect Split on Dashboard → Connections, then add Memtro to your client once. The Split tools appear alongside the tools of everything else you have connected; nothing is configured per connector.

claude plugin marketplace add memtro-org/claude-memtro
claude plugin install memtro@memtro --config api_key=sk-memtro-...

{
  "mcpServers": {
    "memtro": {
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

{
  "mcp": {
    "memtro": {
      "type": "remote",
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

Who may use it

The key identifies the person, and the server offers that person the tools they may use everywhere else in Memtro. A shared Split connection is used by the teams an administrator granted it to, tool by tool, on the Permissions page; a personal connection is used by its owner alone, and its tools that act stay off until the owner allows them. See Teams and permissions and Connectors.

Ask it like this

• › Is paywall_beta killed in production?
• › Which flags in the checkout project are tagged experiment?
• › How many accounts are in the beta_testers segment in Production?
• › Kill paywall_beta in Production.
• › Add accounts 4417 and 5120 to the beta_testers segment in Staging.

Questions

Is this an official Split MCP server?

No. It is Memtro's MCP server with Memtro's Split connector behind it: the connector reads Split through its API with the credentials you connect on Dashboard → Connections, and Split publishes nothing here. The same endpoint carries every other system you connect and Memtro's memory tools.

What does it cost?

Nothing on top of Memtro. One price per organisation by its number of users, from $50 a month for up to 10 users after a 7-day free trial, with every connector, the MCP server, jobs and memory included. You bring your own model keys.

Can it change anything in Split?

Only when you ask, and only what it has been granted: kill a feature flag in one environment, so every user gets its default treatment; restore a killed feature flag in one environment; create a feature flag for a traffic type; change a flag's description, add a tag or set its rollout status; delete a feature flag permanently from every environment; submit a change request to kill or restore a flag, notifying the named approvers; approve, reject or withdraw a change request; add keys (users, accounts) to a segment in one environment; create a segment for a traffic type. Every other tool reads, searches or lists. An administrator grants each acting tool by name per team, and a personal connection acts only after its owner allows it.

Which clients work?

Any MCP client that speaks Streamable HTTP: Claude Code, Cursor, Claude Desktop, OpenCode, Codex, Zed, Windsurf, Cline and the Anthropic and OpenAI SDKs. There is no local process to run; the server is remote and authenticated with your key.

Other MCP servers · Engineering and operations

• GitHub MCP server
• AWS CloudWatch MCP server
• Sentry MCP server
• ServiceNow MCP server
• Okta MCP server
• Twilio MCP server