Install
openclaw skills install clipcatClipcat - TikTok e-commerce video creation skill. Video search, product insights, viral replication, product-to-video generation, breakdown analysis, and video download via Clipcat CLI.
openclaw skills install clipcatUse this skill when you need TikTok e-commerce video creation through clipcat.
Get API key: https://clipcat.ai/workspace?modal=settings&tab=apikeys
This skill is intentionally short. Detailed flags and supported values belong to the CLI itself — always treat clipcat -h and clipcat <subcommand> -h as the primary reference.
This skill is auto-installed by OpenClaw using the declared install spec in the frontmatter above. OpenClaw downloads the platform-specific binary from versioned, immutable URLs under https://static.clipcat.ai/public/cli/vX.Y.Z/ and places it under ~/.openclaw/tools/clipcat/. No remote shell script is executed.
After installation, configure your API key once:
clipcat config --api-key <your-key>
clipcat is the local entrypoint for all Clipcat AI video generation workflows:
clipcat -h to see all commands.clipcat <subcommand> -h to see flags.--pretty when the result is meant for human terminal reading.search — find viral TikTok videos by keyword; supports --region, --sort-by relevance|likes, --time-range any|day|week|month|quarter|half_year, --require-shopsearch_items — search TikTok Shop products by keyword; returns market insights, competitor shops, and product intelligence; supports --region, --offset, --page-token for paginationproduct_detail — get product info by --input <ID or URL>; supports --regionproduct_comment — get product reviews by --input <ID or URL>; supports --region, --sort-rule, --filter-type, --filter-valuereplicate — replicate a viral video with your product images (auto-detects URL type); images via --image (local) or --image-url (URL); local files and URLs can be mixed; supports --model, --duration, --size, --lang, --resolution, --character-idproduct_video — generate video from product images only (no reference video); images via --image (local) or --image-url (URL); local files and URLs can be mixedimage — generate an AI image from a text prompt using GPT Image 2 model; optionally supply up to 5 reference images via --image (local file) or --image-url (URL). Use --aspect-ratio to pick 1:1 (default) / 16:9 / 9:16list_images — list image generation tasks from server; supports --status / --limit / --page filtersbreakdown — analyze a video (script, scenes, music); returns cached result immediately if previously analyzeddownload — download TikTok/Douyin video (returns signed URL); cached results return immediatelyuser_videos — get a TikTok user's video list with analytics (plays, likes, shares, comments, e-commerce cart data); --unique-id required; pass --sec-user-id to skip ID resolution and speed up response; supports --max-cursor pagination and --sort-type 0|1query_task — check status of a task by ID and type (--type replicate | product | breakdown | download | image). Omit --task-id to resume the latest local task.list_tasks — list recent video-related tasks from server (--type required: replicate | product | breakdown | download). Image tasks use list_images.clipcat replicate automatically detects the URL type:
/replicate_from_social (costs 1 extra credit for download)/replicateAlways inform the user about the extra credit before running with a social URL.
clipcat://... strings seen in earlier turns are stable asset references. Pass them verbatim to any --image-url / --character-id flag — never prepend https:// or modify them. See subcommand -h for details.
replicate, product_video, image, and breakdown are async. All four
submit and return immediately with a task ID — they never block.
Typical durations: image ~3 min, breakdown a few minutes, product_video /
replicate 10+ min. Never try to wait synchronously inside a single tool
call — every realistic agent harness has a tool-call timeout (commonly 60s)
that will kill the call long before the task is done. Always go submit → return
→ poll across turns.
~/.clipcat/tasks.json automatically.clipcat query_task --task-id <id> --type <type>. Each
call returns immediately with the current status. Omit --task-id to resume
the latest task. Re-invoke the command across turns (suggested cadence:
~30s for image, ~1-2 min for breakdown / product_video / replicate)
until status is completed or failed.clipcat list_tasks --type <replicate|product|breakdown|download> to
see tasks of a given type from the server.clipcat query_task with no flags automatically reads the latest task from ~/.clipcat/tasks.json and resumes it. No need to remember task IDs.
Trial channels are per-clip pricing; standard channels are per-second pricing.
| Model ID | Duration | Resolution | Notes |
|---|---|---|---|
veo3.1fast | 8s, 16s, 24s | 720p, 1080p | Trial, default. Balanced quality and cost |
veo3.1pro | 8s, 16s, 24s | 720p, 1080p | Trial. Google Veo 3.1 Pro, high-quality variant |
omini_flash | 10s | 720p, 1080p | Trial. Gemini Omni Flash, Google's newest model |
grok_imagine | 10s, 15s, 20s, 30s | 720p | Trial. 9:16 aspect ratio only, longer clips |
sora2_official_exp | 4s, 8s, 12s | 720p | Trial. OpenAI Sora 2 official channel, 9:16/16:9 |
seedance2 | 4-15s (any integer) | 480p, 720p, 1080p | Standard. ByteDance Seedance 2, top quality |
seedance2_fast | 4-15s (any integer) | 480p, 720p, 1080p | Standard. ByteDance Seedance 2 Fast, fast variant |
happyhorse10 | 3-15s (any integer) | 720p, 1080p | Standard. Alibaba HappyHorse 1.0 |
Always check clipcat replicate -h for the current model list.
--lang)en zh fr de ms vi th ja ko id fil es
clipcat -h first if unsure which command to use.query_task across turns to track long-running tasks.X-Amz-* params that break if truncated.--pretty only for human-facing terminal display.