Let AI handle the repetitive work

Attach AI-powered automations to any SOP step — they execute code, call APIs, and browse the web automatically.

AI Agents

Step-level automation

Attach code automations to individual SOP steps. When a run reaches that step, the agent executes it automatically.

Sandboxed execution

Code runs in isolated Docker containers with Node.js, Python, or Bash — safely and with full logging.

400+ app integrations

AI agents can browse the web, call APIs, and work across hundreds of applications using your stored credentials.

How AI Agents works

1

Step 1

Write or generate automation code for any SOP step

2

Step 2

Configure credentials and test in sandbox before going live

3

Step 3

Launch runs and watch agents execute steps automatically with full observability

AI Agents overview
AI Agents in action
AI Agents detail view
Detailed view

Frequently asked questions

AI Agents support JavaScript, TypeScript, Python, and Bash. Code executes in isolated Docker containers (sandboxes) with full logging, so automations run safely without affecting your platform.

The AI Coding Agent (V1.5) uses OpenAI function-calling to generate automation code from natural language descriptions. Describe what you want the step to do, and the agent writes the code — which you can review, test, and refine before going live.

Credentials are encrypted with AES-256-GCM and stored securely. You bind specific credentials to individual SOP steps, so each automation only has access to the API keys, OAuth tokens, or passwords it needs.

Step-level automation is available on Pro and Business plans. These plans include AI builder credits for code generation and AI execution tokens for running automations.

Connect your operations to AI execution.

OKiDO structures your procedures, connects your systems, and makes human + AI execution visible, enforced, and auditable — across your entire organization.