@startanaicompany/techsaac-cli

CLI and Node.js client for the TechSAAC AI Agent Orchestrator.

Install

npm install -g @startanaicompany/techsaac-cli

Setup

export MCP_ORCHESTRATOR_API_KEY='us_...'
export MCP_ORCHESTRATOR_URL='https://tech.startanaicompany.com/api/mcp'

CLI Usage

# List agents
techsaac list-agents --limit 10

# Get agent details
techsaac get-agent --agent-id <uuid-or-name>

# Start / stop / restart
techsaac start --agent-id <uuid>
techsaac stop --agent-id <uuid>
techsaac restart --agent-id <uuid>

# Send message to agent
techsaac send --agent-id <uuid> --message "deploy to staging"

# Create agent linked to a rolez role — context, skills, subagents,
# and container image are all resolved from rolez at every restart.
# Slug must exist in https://rolez.startanaicompany.com.
techsaac create-agent <product_id> "Support Lead" acme-support-lead-eu \
  --rolez_slug support-lead

# Switch (or unlink) the rolez role on an existing agent
techsaac update-agent <agent_id> --rolez_slug premium-support-lead
techsaac update-agent <agent_id> --rolez_slug ''   # unlink

# JSON output
techsaac list-orgs --format json

# Call any API tool directly
techsaac call list_agents '{"limit":5}'

Run techsaac help for the full command list.

Programmatic Usage

import { SaacClient } from '@startanaicompany/techsaac-cli';

const client = new SaacClient({
  apiKey: process.env.MCP_ORCHESTRATOR_API_KEY!,
  url: process.env.MCP_ORCHESTRATOR_URL!,
});

// Or auto-detect from env vars
const client2 = SaacClient.fromEnv();

const result = await client.call('list_agents', { limit: 5 });
console.log(result);

License

MIT