Skip to main content

Overview

The AKOOL CLI is agent-first by design, wrapping the full Open API so developers and AI agents can perform face swaps, generate talking photos, translate videos, and more — straight from the terminal. Built for:
  • Coding agents — Claude Code, Codex, and others
  • CI/CD pipelines — batch media processing, automated workflows
  • Bulk operations — translate 100 videos in one shell loop
  • Custom integrations — wrap it in your own tool

Install

Single static binary, no runtime required. Installs to ~/.local/bin. You only need AKOOL API credentials — see Authenticate below.

Updates

Authenticate

Choose one of the two options below. The first is agent- and CI-friendly; the second is for humans. 1. Environment variables — agents, CI; ephemeral, no file on disk:
2. Interactive login — humans; credentials stored at ~/.akool/config.json:
Enter your Client ID and Client Secret when prompted. Get your credentials at akool.com → API Credentials. Verify your credentials:

Global Options

--json works on any command:
Switch to test environment:

Quick Start

1. Face swap an image:
2. Generate an image:
3. Synthesize speech:
Add --wait to block until the job completes. Use --json for raw JSON output.

Commands

Pattern: akool-cli <command> [subcommand] [options]. Every command supports --help.

Command Reference

login

Configure API credentials interactively. Credentials are stored at ~/.akool/config.json.

credit

Check your account credit balance.

faceswap

Image face swap (Pro V4)

With face enhancement:
With async webhook callback:

Face Swap Plus V4 (multi-face, image + video)

Style options: realistic, beautify, lossless

Video face swap

Query & manage results

Live / real-time face swap

character-swap

Create an animated video from a character image and source video.
With options:
Query result:

image

Text-to-image

Image-to-image

Query result

Upscale or generate variations

image2video

Create a video from a static image.
Premium HD, 10 seconds, 1080p:
With audio soundtrack:
Query & manage:

voice

List voices

Text-to-speech

Query result

upgrade

Tips

  • JSON output: Add --json anywhere for raw JSON, great for scripting:
  • Switch environments: Use -e test to hit the test API:
  • Custom API host:
  • Custom CDN for upgrades:
  • Async tasks: Most generation commands return immediately with a job ID. Use --webhook for server callbacks, or poll with the result subcommand.