Free Open-Source Video Maker · Open Skill 공개 스킬
Free Open-Source Video Maker
Free Open-Source Video Maker은 Digen의 공개 Open Skill입니다. 열어 보면 이 워크플로가 무엇을 만드는지 알 수 있습니다.
VTuber
32 6 21
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/1648313/2026-09-26/u88nur5q.jpg
# Free Open-Source Video Maker ## When to use - The user wants to make an AI video but explicitly does NOT want to spend credits / API money. - The user asks to use free, open-source, or free-tier AI video tools (Gemini/Veo, Pika, Runway, Kling, Luma, Krea, Stable Video Diffusion, AnimateDiff, ComfyUI, HunyuanVideo, LTX-Video, etc.). - The user wants to download free stock/source clips (YouTube/TikTok/Instagram) and stitch them into a video. - The user wants a step-by-step plan, ready-to-paste prompts, or a written guide for making a video with free tools. ## Important constraint (tell the user up front) - The platform's built-in generators (`generate_image`, `generate_video`, `generate_tts`, `generate_music`, `mix_clips`) consume credits. This skill NEVER calls them. It only uses free research, download, and document tools. - This skill does not generate the video pixels itself. It researches the best current free/open-source tool for the user's need, gives exact steps and ready-to-paste prompts, downloads free source material, and saves everything as a document the user can follow. ## Steps 1. Clarify the request: what kind of video (short clip / talking-head / product pitch / short film / music video), how long, what subject/style, and whether the user wants (a) a guide+prompts to run in a free tool, (b) downloaded free stock clips stitched together, or (c) both. If unclear, ask one short round of questions. 2. Research the best free/open-source option for this specific need with `search_web` (use `include_answer=true`, `search_depth='advanced'`). Search for the current best free AI video tool for the user's use case, e.g. "best free AI video generator 2026 no watermark", "free open source text to video model", "Stable Video Diffusion / AnimateDiff / ComfyUI tutorial", "Gemini Veo free tier", "Pika / Runway / Kling / Luma free plan". Gather: tool name, free-tier limits, whether it's open-source, how to access it, and any watermark/credit caveats. 3. Recommend the single best option (or a short ranked list) and explain why it fits — free tier, open-source, quality, ease of use. Be honest about limits (e.g. watermarks, length caps, queue times, needing a GPU for local open-source models). 4. Produce a concrete, ready-to-follow plan: exact steps to use the tool, and ready-to-paste prompts (image prompt + video prompt) written for that tool. If the user wants a talking video, also give a short script and suggest a free TTS option (e.g. Edge TTS, Google TTS, or the tool's built-in voice). 5. If the user wants to assemble a video from free source clips, use `download_video` on YouTube/TikTok/Instagram links they provide (or that you find) to get downloadable clips, and `fetch_video_info` to check a video's stats/duration before downloading. Note: download works for YouTube/TikTok/Instagram only; Facebook/Twitter/LinkedIn are not downloadable — use `fetch_video_info` for those instead. 6. Save the full deliverable (tool recommendation, steps, prompts, script, list of downloaded clips) as a document with `create_document` (name like "Free Video Plan — <topic>", entity_type "documents"). If the user later asks to change it, update it with `update_document` using the returned asset_id. 7. Summarize in chat: the recommended free tool, the key steps, and where the saved document is. Remind the user that no credits were used. ## Notes - NEVER call `generate_image` / `generate_video` / `generate_tts` / `generate_music` / `mix_clips` / `recognize_image` / `recognize_video` / `transcribe_audio` — this skill is strictly zero-credit. - Be honest about free-tier limits and watermarks; don't promise unlimited free generation that doesn't exist. If a tool has a free tier with a watermark, say so and offer a watermark-free open-source alternative if one fits. - For local open-source models (Stable Video Diffusion, AnimateDiff, ComfyUI, HunyuanVideo, LTX-Video), note the hardware requirement (GPU/VRAM) and point to the official install guide.