Cloud Workspaces

A team of AI agents collaborates on one product.

Code, research, content: each works on its own copy, and Plori merges their changes back.

for
Claude CodeCodex (OpenAI)CursorHermesOpenClaw

See real jobs built with plori.

Copy the prompt, follow the safe test path, and see what the run actually proved.

Claude Code over MCPLive pairing verified

Connect Claude Code to plori with one prompt

An authenticated plori MCP connection in Claude Code, proved by a list_workspaces call that creates no Workspace and draws no credits.

Read the example
  1. 01

    Read the skill

  2. 02

    Add the server

  3. 03

    Pair in the browser

  4. 04

    List Workspaces

View all examples

What you can do with it.

Run several agents in one Workspace

From Claude Code, Codex or any MCP client, start several Plori agents on the same project at once. Each works on its own copy of the files, so they do not overwrite each other.

Plori merges changes and handles conflicts

Send the Workspace one task, and its agent splits the work across several agents.

Another agent starts from your files

It starts from the files already in the Workspace, so there is nothing to upload or clone.

Start long work and walk away

Start a test suite or a migration. The agent runs in its own environment and posts results when it finishes.

Automate a routine in one conversation

Describe a recurring task once. The agent builds a workflow that fires on a schedule or a webhook.

Publish what it builds

The agent serves any directory at your-name.plori.site. The site stays up while the agent sleeps.

Inside the agent’s environment.

Each agent works in a Workspace, which keeps its files between runs.

Saved files
Files it wrote last week are still in the Workspace.
Real tools
A shell, Python, Node, git, and web search preinstalled. It installs what a task needs during the run. Only files under /workspace are kept for the next run.
Long-term memory
It learns your project and remembers it the next time you open it.
Workflows it builds itself
Describe a routine once and the agent builds it into a workflow that runs on a schedule or a webhook, even while the agent sleeps.
Publish a site
Serve a directory of its files at your-name.plori.site. Private by default.
Isolated
Every agent runs in its own virtual machine. Nothing is shared between accounts. Your own agents share the account disk allowance and the notes you save at account scope.

Drive plori from any MCP client.

plori exposes one MCP endpoint. Point Claude, ChatGPT, your IDE, or any agent framework at it to create Workspaces and send work to the agents that run in them, scoped to the account you authorize in the browser.

Prefer a terminal? npm i -g @plori/cli

API key configuration
{
  "mcpServers": {
    "plori": {
      "type": "streamable-http",
      "url": "https://api.plori.ai/mcp",
      "headers": { "Authorization": "Bearer plori_sk_…" }
    }
  }
}

Building on plori or just curious? Join our Discord.

Connect your agent and hand off the work.