blitzreels-faceless

📁 blitzreels/agent-skills 📅 7 days ago
10
总安装量
6
周安装量
#30976
全站排名
安装命令
npx skills add https://github.com/blitzreels/agent-skills --skill blitzreels-faceless

Agent 安装分布

amp 6
opencode 6
kimi-cli 6
codex 6
github-copilot 6
gemini-cli 6

Skill 文档

BlitzReels Faceless Video Generation

Generate AI-powered faceless videos from a topic or script.

Cost Warning: Faceless generation consumes AI credits (AI script writing, image generation, voiceover, video assembly). You MUST present a plan to the user and get approval before calling any expensive endpoint.

Quick Start

# Full pipeline: topic → video
bash scripts/faceless.sh --topic "5 productivity tips" --duration 30 --voice rachel --style cinematic

# Dry run — show plan without executing
bash scripts/faceless.sh --topic "5 productivity tips" --dry-run

# From a user-provided script (paragraph breaks = scene breaks)
bash scripts/faceless.sh --script "First scene narration.\n\nSecond scene narration." --duration 60

Required Workflow

Always use plan mode. Before generating, present this plan to the user for approval:

  1. Topic or script — What the video is about
  2. Duration — 10–120 seconds (drives scene count)
  3. Visual style — See references/visual-styles.md or GET /visual-styles
  4. Voice — See references/voices.md or GET /voices
  5. Image model — See references/models.md (default: gemini-2.5-flash)
  6. Video model — See references/models.md (default: kling-2.1)
  7. Output mode — Full video, storyboard-only (images), or images + audio
  8. Captions — Enabled by default; optionally pick a caption style
  9. Plan summary — Present all choices + “this will consume AI credits” warning
  10. User approval → Execute

Input Modes

Mode Flag Behavior
Topic --topic "..." AI writes the script based on the topic
Script --script "..." User provides exact narration. Paragraph breaks (\n\n) become scene breaks.

If both are provided, --script takes precedence.

Scripts

scripts/faceless.sh

Purpose-built end-to-end pipeline. Handles: create project → generate → poll → export → download URL.

faceless.sh --topic TEXT [--duration 30] [--voice rachel] [--style cinematic]
            [--image-model ID] [--video-model ID] [--video-duration SEC]
            [--storyboard-only] [--no-animated-video]
            [--aspect 9:16] [--captions true] [--caption-style ID]
            [--resolution 1080p] [--name TEXT] [--yes] [--dry-run]

Run bash scripts/faceless.sh --help for full usage.

scripts/blitzreels.sh

Generic API helper for ad-hoc calls. Requires BLITZREELS_ALLOW_EXPENSIVE=1 for expensive endpoints.

bash scripts/blitzreels.sh METHOD PATH [JSON_BODY]

Environment Variables

Variable Required Description
BLITZREELS_API_KEY Yes API key for authentication
BLITZREELS_API_BASE_URL No Override base URL (default: https://www.blitzreels.com/api/v1)
BLITZREELS_ALLOW_EXPENSIVE No Set to 1 to allow expensive calls via blitzreels.sh

API Endpoints

Method Path Description
POST /projects Create a new project
POST /projects/{id}/faceless Start faceless generation (expensive)
GET /jobs/{job_id} Poll job status
GET /projects/{id}/context?mode=timeline Inspect generated timeline
POST /projects/{id}/export Start export (expensive)
GET /exports/{export_id} Get export status + download URL
GET /voices List available voices
GET /visual-styles List available visual styles
GET /caption-styles List caption style presets

Faceless Request Body

{
  "topic": "5 morning routine tips",
  "duration_seconds": 30,
  "visual_style": "cinematic",
  "voice_id": "21m00Tcm4TlvDq8ikWAM",
  "captions": true,
  "caption_style": "bold-center"
}
Field Type Required Description
topic string one of topic/script AI generates script from this
script string one of topic/script Exact narration text, paragraph breaks = scenes
duration_seconds number yes 10–120
visual_style string no Style ID (default: cinematic)
voice_id string no ElevenLabs voice ID (default: Rachel)
captions boolean no Enable captions (default: true)
caption_style string no Caption style preset
imageModelId string no Image generation model (default: google-gemini-2.5-flash-image)
videoModel string no Video generation model (default: kling-2.1)
videoDuration number no Per-clip duration in seconds, 3–10
storyboardOnly boolean no Generate images only, skip video/audio (default: false)
generateAnimatedVideos boolean no Set false to skip video animation (default: true)

Generation Modes

Mode Flag What’s Generated
Full video (default) — Images → video clips → voiceover → music → captions
Storyboard only --storyboard-only Scene images only (no video, no audio). Cheapest option for previewing.
Images + audio --no-animated-video Scene images + voiceover + music (slideshow-style, no video animation)

References

  • references/models.md — All image and video generation models with costs and capabilities
  • references/voices.md — All available voice IDs, categories, and descriptions
  • references/visual-styles.md — All available visual style IDs and descriptions

Safety

  • faceless.sh prompts for confirmation before calling expensive endpoints (bypass with --yes)
  • blitzreels.sh requires BLITZREELS_ALLOW_EXPENSIVE=1 for /faceless and /export paths
  • Always present cost/plan to user before generation

Notes

  • Use https://www.blitzreels.com/api/v1 as your base URL. https://blitzreels.com redirects to www, and some HTTP clients drop the Authorization header on redirects.
  • POST /projects/{id}/faceless returns a plan and input_warnings to help you predict what will be generated.
  • Full OpenAPI spec: https://www.blitzreels.com/api/openapi.json

Rate Limits

Plan Requests/min
Free 10
Creator 30
Pro 60
Agency 120