Rilo

Turn a story into a finished video — AI-generated script, voiceover, keyframes, and composition, all in one command.

📖 Docs · 🇺🇸 English · 🇨🇳 简体中文 · 🇩🇪 Deutsch

Rilo turns a story into a finished video — AI-generated script, voiceover, keyframes, and composition, all in one command.

Write your story in plain text. Rilo handles the rest: script generation, narration, visual keyframes, video segments, and final composition — with optional subtitle alignment and burn-in.

Built for creators and teams who need reproducible, high-quality video at scale without manual editing.

Features

• Full pipeline, one command — Story → script → voiceover → keyframes → segments → final video. rilo --project demo --story-file ./story.txt
• Checkpointed runs — Every stage saves artifacts. Resume or regenerate any stage selectively. rilo --project demo --force
• Your models, your control — Choose T2I and I2V models. Override per-model options. Switch models anytime.
• Code-driven pipeline — Deterministic orchestration, checkpointing, and artifact management. Tokens spent on generation, not infrastructure.
• Subtitle alignment & burn-in — Auto-align subtitles to voiceover timing. Burn them into the final video.
• Preview dashboard — Web UI for project management, regeneration, and asset preview. rilo preview
• HTTP API & webhooks — Bearer-token auth, OpenAPI 3.1 spec, webhook subscriptions. Firebase or local.
• Cross-platform — macOS, Linux, Windows. Node.js 22+ and ffmpeg.

Quick Start

Requirements: Node.js 22+, ffmpeg in PATH, and a Replicate API token.

npm install -g @telepat/rilo
rilo settings
rilo --project demo --story-file ./story.txt

Expected outcome:

• A project folder is created under projects/demo/.
• The full pipeline runs through script, voiceover, keyframes, segments, and composition.
• Final video is written to projects/demo/final.mp4.
• Dashboard preview available via rilo preview.

Requirements

• Node.js 22+
• ffmpeg in PATH
• Replicate API token
• macOS, Linux, or Windows

How It Works

Rilo runs a staged pipeline: script generation, voiceover synthesis, shot prompt generation, keyframe rendering, segment generation, and final video composition. Each stage writes checkpointed artifacts so you can resume or regenerate selectively.

Configuration merges CLI flags, environment variables, and ~/.rilo/config.json with schema defaults. The preview dashboard (rilo preview) starts a local API, worker, and Vite React frontend for monitoring and editing.

Using With AI Agents

Rilo provides multiple surfaces for agentic and automated workflows:

• CLI automation — All generation is driven by CLI flags and environment variables. No interactive prompts are required after initial setup.
• HTTP API — rilo preview starts an Express API with full job and project CRUD, asset serving, and webhook endpoints. Bearer-token auth via Authorization: Bearer <API_BEARER_TOKEN>.
• OpenAPI spec — Auto-generated OpenAPI 3.1 spec for schema-driven agent integration.
• Webhooks — Subscribe to job lifecycle events for external orchestration.
• Firebase Functions — Deploy src/api/firebaseFunction.js for serverless API hosting.
• Agent docs — API Reference covers endpoints, auth, and webhooks.

Security And Trust

• API tokens and Replicate credentials are stored in the OS keystore (macOS Keychain, Windows Credential Manager, Linux Secret Service) when available.
• Falls back to an AES-256 encrypted file at ~/.rilo/.secrets if no native keystore is available.
• Environment variables (TELEPAT_REPLICATE_TOKEN, RILO_API_BEARER_TOKEN) take highest precedence and override stored values.
• Preview --expose mode should only be used on trusted networks or isolated environments.

Documentation And Support

• Documentation site
• Quickstart
• CLI Reference
• Configuration Guide
• API Reference
• Troubleshooting
• Repository
• npm package

Contributing

Contributions are welcome. See Development for local setup, build commands, and test workflows.

License

MIT. See LICENSE.