$ show calesthio/OpenMontage
OpenMontage is an open-source, agentic video production platform where an AI coding assistant (Claude Code, Cursor) orchestrates the entire pipeline. It provides a library of 100+ tools (video generation, audio, graphics, enhancement, analysis, avatar, subtitles) behind a scored provider selection layer, and 10+ production pipeline manifests (animated explainer, cinematic, documentary montage, talking head, etc.) that define stage order, approval gates, and tool assignments. The core architecture is a Python framework with a BaseTool class, ToolRegistry for auto-discovery, pipeline loader, checkpoint system, and a Remotion-based video composer. It also includes a Backlot living storyboard server for live production monitoring and an Ink Theater procedural puppetry engine. While primarily an application, it contains data/research elements such as ManimGL/ManimCE educational animation examples covering transformer internals, quantum mechanics, and ML concepts, as well as a CLIP-based corpus builder for documentary montage pipelines.
The LLM identified 6 files worth highlighting for reuse:
Provides the foundational BaseTool class that all tools inherit from. It enforces a consistent interface (execute, dry_run, get_status, estimate_cost) and includes dependency checking, resource profiles, retry policies, idempotency keys, and side-effect declarations. This pattern is highly reusable for any multi-provider tool ecosystem and demonstrates careful design for autonomous agent execution.
Implements rank_providers() which scores providers across 7 dimensions (task fit, output quality, control, reliability, cost efficiency, latency, continuity) with configurable weights. This is a clean, self-contained scoring algorithm that can be extracted and applied to any multi-provider selection problem, demonstrating non-obvious data engineering skill.
Defines production type classification (motion-led vs still-led) and validates cuts against motion requirements. It enforces delivery promises with rules for motion ratios and still-fallback allowances. This is a standalone module that could be reused in any video production pipeline to ensure output matches intended style.
Defines a StockSource Protocol with Candidate and SearchFilters dataclasses, providing a unified interface for stock media adapters. Adapters implement search() and download() methods. This pattern enables pluggable stock sources (Unsplash, Dareful, NARA, Pond5, etc.) and is easily extractable for any project needing normalized media sourcing.
A collection of reusable Remotion components (BarChart, LineChart, PieChart, KPIGrid, TextCard, StatCard, HeroTitle, SectionTitle, StatReveal, ProviderChip, CaptionOverlay, EndTag, etc.) with spring-based animations, staggered entrances, and theme support. These components demonstrate sophisticated React animation patterns and can be extracted for any Remotion-based video project.
Manages project state with checkpoint read/write, validation, prerequisite enforcement, approval gates, and history archiving. It provides a robust state machine for multi-stage pipelines that can be reused in any workflow system requiring structured progress tracking and gate approvals.