Connected-App Workflow Sources

The workflow selection covers Higgsfield, Meta, browser and computer use, Slack, Context7 and the instructor's own Resolve work. Shorts and Reels research added further examples. Public video evidence comes from captions, descriptions, metadata and timed captions. The research did not include watching every video frame. Complete reports and the retrieved primary MapQuest record informed the narration, with claims kept within those sources.

View and engagement counts below are research snapshots. Popularity does not establish adoption or a successful workflow.

Creative work with Higgsfield

Duncan Rogoff's Higgsfield creative workflow starts with an ordinary lens-cleaner photograph, then develops a brief, research and shared references for coordinated images and video concepts. The roughly 2,660-word caption record describes the demonstration. The workflow was not reproduced locally, and the source does not establish that ads were published. The retrieved count was 17,102 views. The source includes an affiliate disclosure. The case was selected for its concrete starting material and useful handoff of several assets.

Higgsfield's motion-promo Short describes a nature-reserve promo using imagery, a route map and a landmark, followed by a render request and a revision. The official vendor promo is a 73-second Short, with 99,987 views and 2,030 likes in the retrieved record. Captions describe requests. An editable After Effects handoff or client delivery was not verified, and no specific model claim entered the lesson. The recent view count and connected request sequence supported the creative workflow example.

Advertising with Meta

Felipe Vergara's Meta workflow moves from account information and strategy through a missing-ID failure and a narrower request to campaign structures in Ads Manager. Some targeting still needs manual work. The complete Spanish timed captions for 24:40 through 30:06 describe four campaigns; the retrieved record showed 82,891 views. The source does not establish activation, successful complete targeting or later performance. Limitations from the May tutorial are not permanent October product facts. The case was selected because it includes a result and a correction.

The Meta connection lesson demonstrates two successful setup routes. A small read-only account task remains optional follow-up practice.

Browser journeys and account information

Elie Steinbock's Browser MCP examples cover a new onboarding feature tested through a browser journey and a signed-in fantasy-football team combined with separate search information for recommendations. Complete captions were retrieved from this 2025 source, with 8,756 views. The client shown is Cursor/Claude with Browser MCP. The lesson does not claim that the source originally used Claude Code or Browser Use. An exhaustive regression test or completed transfers were not established. The specific app change and account facts explain what a browser connection contributes.

Software documentation with Context7

Jessica Wang's Context7 Short describes retrieving documentation during React and TypeScript work. The complete caption record has 181 words; the Short lasts 52 seconds and had 43,297 views in the retrieved record. The client is Warp, and a completed clone was not verified. Research corroborated Context7's documentation role. The case supports a bounded documentation step, with source credit.

Video production with Resolve

Nick Automates' Resolve Short describes timeline work and generated video material. The 33-second Short had 53,716 views and 1,299 likes in the retrieved record. Captions provide an overview, rather than independent execution of each feature. Claims about a free editor, fully automatic production and specific models were excluded. The source signals interest in a workflow supported more strongly by the instructor's own production work.

The instructor's Resolve production assembled earlier Skills lectures from recordings, narration and timelines. The 1 October journal records all 15 timelines, with no empty media frames or missing media. The 2 October journal records 189 clips repointed and a self-contained handover. Primary evidence checked in the course work supports Blackmagic as the official MCP publisher. The actual production job and handover provide stronger evidence than the public Short alone.

An everyday route with MapQuest

Nate Herk's MapQuest Short describes Claude and MapQuest MCP planning a New York day with breakfast, somewhere to work, three sights and dinner, then ordering the stops. The complete caption and description record has 206 words. The 45-second Short had 11,798 views and 269 likes in the retrieved record. The title and channel identify Nate Herk. MCP and Claude are explicit. The Claude Code client was not verified. No free-transaction claim entered the lesson. The case adds an everyday job with a clear input, action and result.

Slack as a possible workflow

A named project thread could supply decisions and open questions for drafted project tasks with source links. This is a proposed use, rather than a locally recorded execution. Research did not establish a strong recent Slack Short. The older Gavin LinkedIn/Loom claim remains background. Slack was selected for useful coverage, with possibility language.

Other sources considered

These sources informed selection but did not become additional headline cases.

• Higgsfield ads Reel: verified as a 1 October Reel, with 10,439 likes and 7,746 comments. Plays were unavailable. Its caption names ChatGPT/GPT, which would require a detour to avoid confusing the Claude client. Most comments ask for a link. The research supports discovery, rather than adoption or results.

• Higgsfield computer-use Reel: verified as a 30 September Reel, with 3,962 likes and 814 comments. Its context is Codex/ChatGPT, rather than Claude. It was excluded from narration. The lessons already explain computer use by function.

• Sam Piliero's Meta Short: a 42-second Short with 4,381 views. It claims image, copy and campaign duplication, but research did not establish destination readback. Felipe's fuller sequence was preferred for teaching. Sam's Short was not labelled as high views.

• TopView Short: 36,000 views, with product-input and workspace promotional claims. Generated assets were not inspected. It was excluded to avoid another tool catalogue and an unsupported complete outcome.

• Firecrawl, Playwright and Context7 Short: 19,600 views, with a tool list rather than a complete browser job. Elie's substantive browser case was preferred.

• A monster-truck Short had 40,000 views, but its music and exclamation caption did not establish a workflow, so it was excluded.

• Colin Elefante-Smith's X post: a first-person claim about editing 28 minutes of footage with Higgsfield B-roll and sound, with 211 views and two likes. The instructor's Resolve job provides stronger evidence of actual work. No claim about hours or days saved entered the lesson.

• Dylan's X post: a claim about 30 static ads and a Skill, with 2,244 views and 24 likes. Duncan's complete input, steps and handoff explain the mechanism more clearly. The post does not establish a widespread trend.

• Filip Hric's X post: 41,663 views and 501 likes, with Playwright MCP versus CLI and Jev Trello timing claims. The cost and speed comparison was not independently reproduced and was not used as fact. The tool-choice lesson does not borrow the 98 percent claim.

Official setup references and the recorded Meta boundary

For the provider's explanation of Tool Search, use Claude Code MCP guidance. Detailed tool definitions can load when needed. Returned content and conversation history still occupy context.

Higgsfield's integration help is the primary route reference. The researcher verified the recommendation to use the CLI with Claude Code.

The Meta connection lesson shows the official endpoint https://mcp.facebook.com/ads configured through a desktop custom connector and a local project MCP. Human authentication succeeds, the local list shows 97 tools, and a fresh desktop Code conversation reports tools loaded. No account or campaign read, Ads Manager comparison or Claude Ads Skill installation is shown. Follow current official setup instructions for your own connection rather than importing commands from the May tutorial.

How the research was gathered

The research used YouTube Data API searches, captions, timed captions and Shorts status as a vidIQ-equivalent route, Bright Data structured public Instagram profile and Reel readbacks, and X/Grok public-post research. vidIQ was not found in the checked runtime. No vidIQ credits or outlier score were used. Instagram plays were unavailable. These are provenance limits for the source record, rather than claims that each tool or workflow was reproduced.