skills/dumplingai-cli/SKILL.md
When the user wants to find, inspect, or execute a DumplingAI capability or provider endpoint from the terminal. Also use when the user mentions `dumplingai`, `catalog search`, `catalog details`, `dumplingai run`, `balance`, `usage`, `transactions`, `view-config`, or wants to test a DumplingAI-powered workflow without wiring a direct vendor SDK. Use this whenever execution, inspection, or account-level CLI checks should happen through DumplingAI instead of a raw API integration. For choosing the right capability or endpoint first, see discovering-dumplingai-apis. For narrower job-based entry points, see dumplingai-web-research, dumplingai-scraping-extraction, dumplingai-transcripts-media, and dumplingai-seo-data.
npx skillsauth add dumplingai/cli dumplingai-cliInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
The dumplingai CLI is a thin terminal interface for DumplingAI's Unified API Platform under /api/v2.
dumplingai catalog search <prompt>
dumplingai catalog details <type> <id>
dumplingai run <type> <id> --input '<json>'
dumplingai balance
dumplingai usage
dumplingai transactions
dumplingai catalog search "<job>".dumplingai catalog details <type> <id> before execution.dumplingai run..dumplingai/ and inspect them incrementally.dumplingai catalog search works best with short keyword queries. Prefer phrases like google search, scrape page, youtube transcript, keyword ideas, or firecrawl scrape.
If a search comes back empty, shorten it further and remove filler words. Do not assume long natural-language prompts or suffixes like capability and endpoint will improve recall.
Write large outputs to .dumplingai/:
dumplingai catalog search "google search" > .dumplingai/catalog-search.json
dumplingai catalog details capability google_search > .dumplingai/google-search.json
dumplingai run capability google_search --input '{"query":"latest TypeScript release"}' > .dumplingai/result.json
Then read incrementally:
head -40 .dumplingai/result.json
rg '"error"|"results"|"output"' .dumplingai/result.json
For common task families and search phrases, see references/catalog-domains.md.
See rules/safety.md for content trust and credential handling rules.
data-ai
When the user wants to turn a YouTube video into a blog post, article, outline, newsletter draft, or transcript-based written summary. Also use when the user mentions repurposing a video, extracting a transcript, rewriting a YouTube talk into a post, or expanding a creator video into long-form content. Use this whenever the workflow should start with transcript extraction and optional source verification through DumplingAI capabilities.
testing
When the user wants social media content, post variants, threads, captions, hooks, or channel-specific copy for X, LinkedIn, Instagram, TikTok, or YouTube Shorts. Also use when the user wants to turn a topic, campaign idea, product page, article, or transcript into social posts, or asks for platform-specific repurposing instead of one generic draft. Use this whenever research, source extraction, and writing should be routed through DumplingAI-powered capabilities.
development
When the user wants web research, search results, cited sources, or fast topic discovery through DumplingAI instead of manual browsing or a direct search API integration. Also use when the user mentions Google search, SERP results, web search, search for sources, topic research, competitor research, answer sources, Serper, Perplexity, or wants to compare what the web says about something. Use this whenever the job is to search first, then inspect or synthesize results. For scraping a specific page or site after discovery, see dumplingai-scraping-extraction.
data-ai
When the user wants a transcript, captions, spoken-text extraction, or media-to-text workflow through DumplingAI. Also use when the user mentions YouTube transcript, TikTok transcript, transcribe this video, pull captions, extract text from audio, transcript from URL, or wants to inspect the contents of a video before writing or summarizing from it. Use this whenever the job is to turn video or audio into text first. For turning a YouTube transcript into a finished article, see youtube-to-blog-post.