darkzOGx/youtube-automation-agent

What it solves

AgentTube is an open-source AI agent designed to automate the entire lifecycle of a YouTube channel. It removes the manual effort of researching topics, writing scripts, generating audio and visuals, assembling videos, and managing publishing and analytics, while maintaining human-in-the-loop control for quality and rights management.

How it works

The system uses an "Autonomous Channel Operator" that takes high-level objectives, audience targets, and content pillars to create editorial plans. It then executes a production pipeline that integrates with various AI providers (Gemini, OpenAI, etc.) for text, image, and video generation.

Key operational components include:

• Review Studio: A human-review gate where users must approve factual claims, media rights, and final content before scheduling.
• Scene Repair Studio: Allows for granular editing of individual scenes (narration, visuals, timing) without regenerating the entire video.
• DarkzSEO: A discoverability adapter that audits content packaging for search and AI-engine optimization.
• Analytics Loop: Captures YouTube performance data (CTR, retention curves) to generate learning recommendations that the operator can approve to refine future content strategies.
• Engagement Studio: Syncs comments to identify audience-requested ideas and draft replies in the channel's voice.

Who it’s for

Content creators and channel managers who want to scale video production using AI agents while retaining strict editorial oversight and data ownership (self-hosted).

Highlights

• End-to-end automation: Handles everything from trend research to publishing and audience engagement.
• Approval-first workflow: Prevents autonomous publishing by requiring human sign-off on factual claims and media rights.
• Provider flexibility: Supports multiple LLMs and video/image generation engines (e.g., Kling, Wan, Gemini, OpenAI).
• Scene-aware analytics: Maps YouTube retention curves directly to specific video scenes to identify exactly where viewers drop off.
• Growth experiments: Built-in tools to rotate and test title/thumbnail variants to scientifically prove growth recommendations.
• Shorts repurposing: Automatically converts approved long-form videos into short-form drafts based on the scene timeline.