Doodle Cast Creator — AI Video Production Studio

02 / 12

The Problem

Traditional video production requires a team.

Every episode needs coordination across multiple specialists, weeks of effort, and thousands of dollars.

Writer

Animator

Voice Actor

Video Editor

Sound Designer

Music Composer

Thumbnail Artist

Social Media Manager

Publishing Coordinator

Analytics Specialist

Or… one tool.

03 / 12

The Solution

Doodle Cast Creator

An end-to-end AI production studio that transforms a single idea into published, multi-platform video content — episodes, shorts, and podcasts.

0

Lines of Code

0

API Endpoints

0

AI Models

0

Integration Tests

An AI-native production studio that replaces an entire content team.

04 / 12

The Pipeline

From idea to YouTube in 10 steps.

Every stage of episode production — handled by a unified interface with AI assistance at every step.

STEP 01

Creative Director

Three AI personas independently brainstorm concepts, then debate their merits. Grok evaluates each pitch with live web search and selects a winner.

STEP 02

Script Writer

Choose your LLM and generate a full episode script with structured clips, character dialogue, and image prompts. Trained on an evolving show bible.

STEP 03

Voice Readout

Every character speaks in a distinct synthesized voice via ElevenLabs. Play through the full episode readout to check pacing and dialogue flow.

STEP 05

Scene Generation

Photorealistic AI images for each clip, informed by character references, location images, and scene descriptions. Start and end frames for smooth video generation.

STEP 06

Video Timeline

A DaVinci Resolve-style timeline with start/end frames, status badges, and a composite episode preview player that sequences all clips in real time.

STEP 08

Audio Mix

Four-track audio editor: video audio, voice dialogue, sound effects, and music. Each track with independent volume control and keyframe automation.

STEP 09

Render Engine

One-click episode render. Trims clips, applies the full audio mix with ffmpeg filter graphs, concatenates everything, and uploads to Google Drive.

STEP 10

Publish to YouTube

AI-generated thumbnails for A/B testing, AI-written metadata, and direct YouTube publish with real-time upload progress.

05 / 12

Batch Content

20 shorts. 5 platforms. One click.

A parallel 7-step pipeline for producing batches of vertical short-form content — each with unique characters, dialogue, AI video, voice synthesis, and music.

Step 1 — Idea Lab

Step 2 — Script Writer

Step 3 — Start Images

Step 4 — Video Generation

Step 5 — Voice Studio

Step 6 — Render

Step 7 — Publish & Schedule

06 / 12

Audio Production

Podcast Creator. Same characters, new format.

A complete 7-step audio-only pipeline. Generate conversational podcast episodes where the show’s characters discuss real topics — with multi-voice dialogue, AI cover art, and direct publishing.

PODCAST

Idea & Script

Three modes: AI-recommended pitches with Grok scoring, bring your own idea, or Research & Debate with live web research. Scripts include multi-character dialogue with ElevenLabs voice directions and configurable target length.

DIALOGUE

Multi-Voice Generation

ElevenLabs Text-to-Dialogue API generates natural conversations between multiple characters in a single audio stream. Each character maintains their unique voice profile with proper conversational pacing. Act-based batching for long episodes (15+ minutes).

PUBLISH

Preview & Download

Full script preview with character avatars, auto-generated intro/outro music, AI cover art variants, and publish-ready audio output. Perfect for building a complementary audio feed alongside the video channel.

06b

v0.19 – v0.20

News Desk Segments. Late-night news comedy, reimagined.

Extensible segment template system with comedy-first scripting. The Fire Hydrant Gazette segment brings full news-desk format with topic-based workflows, guest correspondents, and OTS news graphics.

📰

Segment Templates

Extensible segment types with format rules, comedy mechanics, joke formulas, voice profiles, guest arcs, and layered bible system. Code-canonical, upserts on boot.

🎬

OTS News Graphics

Character-positioned over-the-shoulder graphics. Rusty sits left, OTS right. Oreo sits right, OTS left. Configurable position/size per segment type. Live preview with adjustable sliders. Composited in final render via ffmpeg.

🎧

Audience Reaction Track

Comedy segments get per-template audience reactions — laughter, applause, gasps. Generated via ElevenLabs, mixed into all render pipelines.

📡

Discord Auto-Announcements

New episodes, shorts, gazette articles, and podcasts automatically posted to the fan Discord server. Non-blocking hooks — publish failures never block the response.

🎤

Per-Segment Intro Clips

Each segment type can have its own intro video. Appears in timeline, included in HLS, VPS fallback, and DaVinci export render pipelines.

🖼

Scene Image Enhancements

Gallery picker on empty scene slots. Drag-to-copy images between clips in the timeline. Per-clip topic image toggle (show/hide). Image history with undo.

07 / 12

Distribution

Publish everywhere. Automatically.

OAuth-connected to all five platforms. AI-generated metadata per platform. Scheduled auto-publisher with email alerts for missed posts.

YouTube

Facebook

Instagram Reels

TikTok

X / Twitter

Discord

Auto-publisher polls every 60 seconds. AI recommends optimal posting times based on channel analytics.

08 / 12

Beyond Entertainment

The pipeline is the product. The content is yours.

The same AI production pipeline that creates animated YouTube episodes can be adapted for commercial video content at scale.

🎬

Product Ad Campaigns

Generate dozens of video ad variants from a single brief — different angles, hooks, and CTAs. A/B test creative at scale without a production budget.

📡

Social Media Operations

Consistent branded content across 5 platforms with character mascots as brand ambassadors. Scheduled publishing with AI-optimized timing.

🏠

Real Estate & Property

Transform property photos into cinematic walkthrough videos with narration, music, and multi-platform distribution.

🛒

E-commerce Product Videos

Turn product photography into engaging video content with AI-generated scripts, voiceover, and motion — at a fraction of traditional cost.

🎓

Educational Content

Multi-character explainer videos at scale. Complex topics made accessible through AI-scripted dialogue between distinct personas.

💼

Agency White-Label

Content volume that would normally require a 10-person team, produced end-to-end by AI. Scale output without scaling headcount.

09 / 12

Why This Is Different

Every other tool solves one piece.
This solves the whole puzzle.

Runway / Veo

Generate a 5–10 second clip

You still need everything else

InVideo / Pictory

Assemble stock footage into a video

Not your content. Not your style.

HeyGen / Synthesia

AI presenter reads your script

Corporate talking heads. Not cinematic.

Showrunner

Generate animated episodes

Animation only. No publishing pipeline.

DC Creator

Idea → Script → Voice → Scenes → Video → Audio → Render → Publish

The complete pipeline.

10 / 12

Architecture

Hybrid cloud + local GPU.

Cloud APIs deliver the highest-quality generation. A local RTX 5090 provides open-source alternatives for free iteration. The VPS orchestrates everything.

☁️

Cloud APIs

Gemini · Veo · Grok · Claude
ElevenLabs · YouTube · Drive
Facebook · Instagram · TikTok · X

🖥️

VPS Orchestrator

Node.js · Express · SQLite
22 routers · 11 services · ffmpeg
302 API endpoints · Auto-publisher

🖌️

Local GPU (RTX 5090)

WAN 2.2 · LTX 2.3 · Z-Turbo
Ollama (Llama, Gemma, Qwen)
Faster-Whisper · NVENC · ComfyUI

11 / 12

Engineering

The April 2026 Refactor.

Production-grade architecture. Not a prototype. Not a demo. A system that runs every day in production.

17,744 lines

441 lines

server.js

-97.5% reduction

22 tests

114 tests

Integration test suite

5x coverage increase

1 monolith

34 modules

22 routers + 11 services + middleware

Fully modular architecture

0 security layers

Hardened

Helmet, rate limits, injection prevention

GitHub Actions CI/CD

The refactor caught 9 serious bugs that the original test suite had missed. Every test runs on a fully fresh database.

12 / 12

What’s Next

This is what AI-native content production looks like.

One tool. Zero production team. A complete AI-native studio that runs in a browser.

See the Output ↗ Technical Deep Dive →