Skip to content

RunAgents Copilot

Copilot lets you manage the entire RunAgents platform in natural language. Describe what you want to deploy, and Copilot figures out the steps, proposes the resources, and creates them after your confirmation.

Console Copilot

Open the copilot panel from the ? button in the console sidebar on any page.

You: deploy a Stripe customer lookup agent

Copilot: I'll set up the following:
  1. Register tool: stripe (https://api.stripe.com, OAuth2)
  2. Deploy agent: stripe-lookup with gpt-4o-mini

  [Confirm] [Skip]

Reads only (list tools, check runs, view billing) execute instantly. Mutations (create tool, deploy agent, approve action) always show a confirmation card before executing.

Configure the model in Settings > Copilot. Requires an OpenAI API key.

CLI Copilot

Interactive terminal session — ideal for developers who prefer the command line.

Install

curl -fsSL https://runagents-releases.s3.amazonaws.com/cli/install.sh | sh

Configure

runagents config set endpoint https://your-workspace.try.runagents.io
runagents config set api-key YOUR_API_KEY
runagents config set namespace default

Start a session

runagents copilot

RunAgents Copilot v1.1.1 — type "exit" to quit, "help" for commands

> deploy this folder as billing-agent

Analyzing source files in ./ ...
  Detected tools:    stripe (https://api.stripe.com)
  Detected model:    gpt-4o-mini (openai)
  Entry point:       agent.py

I'll create:
  1. Register tool: stripe
  2. Deploy agent: billing-agent

[Confirm? y/n] y

  ✓ Tool created: stripe
  ✓ Agent deployed: billing-agent (Running)

> what tools do I have?

You have 3 registered tools:
  • echo-tool (Internal, Restricted posture)
  • stripe     (External, OAuth2)
  • slack      (External, API Key)

Quick commands

What to type What happens
deploy this folder as agent-name Analyzes current directory and deploys
deploy draft <id> as agent-name Deploys from a saved draft
what tools do I have? Lists registered tools
list my running agents Shows agents and status
approve action <id> Approves a blocked action
check my billing Shows plan status and usage

Check readiness

runagents copilot doctor

Runs local + API connectivity checks and reports what's configured.

Single-shot commands

Don't want an interactive session? Use copilot chat for one-off questions:

runagents copilot chat "how do I register a Stripe tool?"

Project session memory

The CLI remembers context within a project directory. Create .runagents/project.json to persist agent name and tool preferences across sessions:

{
  "agent_name": "billing-agent",
  "preferred_tools": ["stripe", "slack"],
  "default_model": "openai/gpt-4o-mini"
}

What's Next