Plori MCP server

What it does

Plori is a remote MCP server that exposes cloud AI agents, each with a persistent disk, CLI tools, and memory. It lets MCP clients create, run, and manage agents and multi-step workflows without installing anything locally. Intended for developers who want to orchestrate cloud agents from their own tools.

• Invoke a cloud agent from an MCP client like Claude Code or Cursor and retrieve its response without managing local infrastructure.
• Define, edit, and execute multi-step workflows on a cloud agent, then poll per-step execution status and output payloads.
• Schedule deferred agent runs and handle human-in-the-loop approval queues from within an MCP-enabled development environment.

Tools this server exposes

Tool	Does
list_agents	Lists all agents on the account.
get_credits	Returns the remaining credit balance for the account.
get_usage	Returns usage data for the account.
get_disk	Returns disk usage information for the account.
list_connections	Lists third-party OAuth provider connections with status, authorization, expiry, and configured scopes.
list_workflows	Lists all workflows, optionally filtered by agent ID or the unassigned bucket.
get_workflow	Reads a workflow with its step projection pinned for execution.
get_workflow_version	Reads the full definition of a specific workflow version.
edit_workflow	Edits a draft workflow under compare-and-swap semantics.
create_workflow	Creates a new workflow, optionally associated with a specific agent.
run_workflow	Runs a built workflow as a real, billed execution.
list_workflow_executions	Lists recent executions of a workflow.
get_workflow_execution	Polls a single workflow execution's status, timing, credits, and per-step input/output payloads.

Before you enable it

An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.