--- title: "Is there a Claude Code skill for UGC-style talking video ads?" description: "Yes, Wireflow's UGC ad skill for Claude Code turns a product photo into a 6 second talking video ad by an AI creator, with sound, for about 106 credits." updated: 2026-10-28 --- Yes, Wireflow's `ugc-ad-generator` skill makes a UGC-style video ad from one product photo inside Claude Code. It writes a hook and a short spoken line, describes an AI creator, makes a selfie-style still of that creator holding your product, then turns the still into a 6 second vertical clip where they say the line, with sound. The render runs as a published Wireflow app: Seedream 5 Lite makes the still and [Veo 3.1 Lite](https://www.wireflow.ai/models/veo-3-1-lite) animates it with voice. You get the still and the MP4 in `./wireflow-outputs/`, plus a reminder to label the creator as AI. Without a Wireflow connection, you still get hooks, lines and creator briefs you can hand to a real creator. Skill files and examples are in [the skill's folder on GitHub](https://github.com/wireflowINC/skills/tree/main/skills/ugc-ad-generator). ## How do I install the UGC ad skill in Claude Code? Install the Wireflow plugin, which adds every Wireflow skill and the Wireflow connector: ```text /plugin marketplace add wireflowINC/skills /plugin install wireflow-skills@wireflow ``` Claude Code shows the plugin's details first, so you can review it and choose its scope. Then run `/mcp`, choose `wireflow` and sign in with your Wireflow account. The skill shows up as `/wireflow-skills:ugc-ad-generator`, and Claude also picks it up from plain requests. To add only this skill, use the skills CLI, or copy the `skills/ugc-ad-generator` folder into `~/.claude/skills/`. Then connect Wireflow once and run `/mcp` to sign in, with no API key to copy. ```bash npx skills add wireflowINC/skills --skill ugc-ad-generator claude mcp add --transport http wireflow https://www.wireflow.ai/api/mcp ``` ## What do I tell Claude to make a talking video ad? Give it a product photo at an `https://` URL and say who you sell to: ```text Make a UGC ad for my canvas tote: https://example.com/tote.jpg Audience: students and commuters. TikTok, casual, no discount. ``` It asks what the product is, who buys it, the one true benefit a viewer can see, and where the ad runs, such as TikTok, Reels or Stories. Offers and prices go in only if they are real. Requests for a "talking head ad", a "creator-style ad" or "TikTok UGC" also start the skill. ## How does it write the script and the creator? It fits one hook and one proof into six seconds: the hook in the first 2 seconds (a reaction, a confession or a "you know when..." moment), then one concrete, true detail in the next 4, about 12 to 18 words in all. It writes two or three options for you to pick from, then describes a plausible everyday creator who matches your buyer: age range, style, setting, light and mood. This is the creator and line from the skill's recorded test on 8 October 2026: ```text Creator: A friendly woman in her late 20s with curly dark hair and a denim jacket, sitting on a sofa in a bright apartment with plants, relaxed and smiling, holding the tote bag up beside her face. Line: "Okay, this tote fits my laptop, my gym clothes and my lunch. Finally, one bag." ``` ## What files do I get? Two files per ad: a selfie-style still and the talking clip. | Stage | Model | Output | | --- | --- | --- | | Selfie still | Seedream 5 Lite, 9:16 | JPEG, 1920x3416 in the recorded test | | Talking clip | Veo 3.1 Lite, 9:16, 6 s, 720p, audio on | MP4, 720x1280 | Watch the clip before you use it and check the line was spoken as written, because the model sometimes paraphrases. The recorded test had not checked the spoken words or lip sync yet. Add captions and music in an editor, and see the test's still and clip on the skill's [examples page](https://github.com/wireflowINC/skills/blob/main/skills/ugc-ad-generator/examples/EXAMPLES.md). ## Is the creator in the ad a real person? No, the creator is AI-generated from a written description. Label the ad as AI-generated where your platform asks for it, and never present it as a real customer testimonial. The skill will not imitate a real person, a celebrity or a specific influencer, will not use someone's face without consent, and will not write fake reviews or results. ## How many credits does one video ad cost? About 106 credits per ad at Wireflow's published prices on 8 October 2026: 10 for the still and 16 per second for 6 seconds of video. | Step | Credits | | --- | --- | | Hooks, lines, creator brief, quote | 0 | | One ad (still and 6 s talking video) | About 106 at published prices, quoted as up to about 225 until re-measured | | Another take | The same again, only with your yes | The only test run so far showed a 224 credit drop while other jobs were spending on the same account, so the skill quotes about 106 and warns that a run can cost more. Video needs credits on your Wireflow account, and video models are on paid plans: see [pricing](https://www.wireflow.ai/pricing) for plans and credits. ## Does the UGC ad skill work in Cursor, Codex or Claude Desktop? It is built for Claude Code, and the skills CLI also installs it in Cursor, Codex and other agents that read `SKILL.md`. Rendering then needs that agent connected to the Wireflow MCP server, and the skill's README says Wireflow has not tested those agents itself. In Claude.ai and Claude Desktop, add `https://www.wireflow.ai/api/mcp` as a custom connector; saving files needs a shell, so there you get the video link instead. Connection steps for each agent are in [Use Wireflow from your agent](/docs/use-wireflow-from-your-agent). Every Wireflow skill is listed on the [agent skills page](/docs/skill). For static ad images in three sizes, use the [ad creatives skill](/docs/skills/ai-ad-creative-generator). To run these ads for many clients from one reusable workflow, see [AI UGC ads for agencies](https://www.wireflow.ai/features/ai-ugc-ads-for-agencies). Last updated: 2026-10-28 --- Documentation index: fetch https://www.wireflow.ai/llms.txt for the full list of Wireflow docs.