Pitch your Startup, App or Hardware or post a Startup Event or Startup Job
Company profile:
Startup name: RefVid
Tagline: Same product, same face, in every AI video you generate
Elevator Pitch: AI video models are good at motion but bad at keeping your product or your character the same from one clip to the next: the label changes, the logo melts, the face drifts. RefVid fixes that with reference-to-video. You save a product or a character once from 1-10 photos (or paste a Shopify or Amazon product link), describe a scene or pick a template (360 turntable, unboxing, UGC ad, hero reveal), and every 3-30 second video keeps the same label, packaging, face and outfit. It runs on Kling 4 by default and offers 9 models in all (Seedance 2.5, Wan 3.0, Vidu Q4 and more). Every new account gets a free 5-second video, no card.
Target Market: Small e-commerce brands, Shopify and Amazon sellers, and creators who need consistent product or character videos
How will you make money?: Credit packs from $9.90 and monthly plans from $10/month; credits are held before a generation and refunded automatically if it fails.
How much capital have you raised?: None
Website: https://refvid.com
City/Country: Dublin, Ireland
AI-assisted summary:
RefVid is a web-based AI video tool that turns reference photos of a product, person, pet, or mascot into short videos designed to keep that subject visually consistent across clips. (Source: https://refvid.com/about) Its core relevance is practical: it targets the continuity problem that makes it difficult to use general-purpose AI video generation for repeatable product creative, recurring presenters, mascots, and short narrative series. (Source: https://refvid.com)
Problem and target users
RefVid is aimed at users who need the same product, logo, face, outfit, or character to remain recognizable while changing scenes, camera angles, hooks, and formats. (Source: https://refvid.com) The company specifically presents use cases including Shopify and Etsy product-page clips, ads, recurring characters, product videos, ad variations, and stories or animation. (Source: https://refvid.com)
The target customer appears to be an individual creator, e-commerce seller, small brand, or production-oriented marketer that has reference imagery but does not want to re-upload and re-specify a subject for each new generation. RefVid says users can save a subject from one to ten images, then reuse it in subsequent prompts and templates. (Source: https://refvid.com/about)
Product and solution
Users upload reference photos, describe a scene or choose a template, select video settings, and generate a clip; RefVid says its workflow supports videos from three to 30 seconds in vertical, square, or widescreen formats, at 720p or 1080p, with optional sound depending on the selected model. (Source: https://refvid.com/about)
- RefVid says it can use one to ten photos per subject and place up to three saved subjects in a single shot, such as a person holding a particular product. (Source: https://refvid.com/about)
- The product offers scene photos, which allow users to provide an image of a shop, kitchen, or set as the intended visual environment. (Source: https://refvid.com)
- It supports continuation between clips, with the stated goal of making a sequence read as one story rather than as isolated generations. (Source: https://refvid.com)
- RefVid defaults to Kling 4 and also exposes other third-party video models, including Kling 3.0 Omni, Seedance 2.5, Wan 3.0, Vidu Q4, Gemini Omni Flash, PixVerse C1, Happy Horse 1.1, and MiniMax H3. (Source: https://refvid.com/terms)
The company’s terms make clear that RefVid is an orchestration and workflow layer over third-party generation models, rather than a claim to operate its own foundational video model. (Source: https://refvid.com/terms) It also states that outputs can vary and are not guaranteed to match a user’s references or prompt exactly, an important qualification for teams considering AI-generated product or brand assets. (Source: https://refvid.com/terms)
Business model, pricing, and traction signals
RefVid uses a credit-based monetization model with pay-per-video purchases, one-time credit packs, and recurring subscriptions. (Source: https://refvid.com) The site lists one-off video pricing of $2.99 for up to five seconds, $4.99 for up to 10 seconds, and $6.99 for up to 15 seconds; it also lists monthly plans beginning at $10 and credit packs beginning at $9.90. (Source: https://refvid.com/about)
The product offers a first five-second video free without a card, says unused credits do not expire, and states that credits are automatically returned for failed video generations. (Source: https://refvid.com/about) RefVid also says commercial use is allowed for free and paid videos, while users retain the rights they have in uploads and, as between the user and RefVid, in generated videos, subject to model-provider terms and applicable law. (Source: https://refvid.com/about) (Source: https://refvid.com/terms)
Public company information is limited: RefVid identifies its operator as Louis Chi in Ireland, and its legal notice lists an Irish address. (Source: https://refvid.com/legal-notice) The reviewed public sources did not provide independently verifiable funding, revenue, customer-count, or enterprise-customer disclosures.
Expert take
RefVid’s positioning is sharper than a generic prompt-to-video tool: visual consistency is a concrete workflow problem for product marketers and creators producing a series rather than a one-off clip. The combination of saved references, multi-subject scenes, templates, and clip continuation could make the product useful for lightweight creative iteration.
The main strategic risk is platform dependence. RefVid’s end-user experience may be differentiated, but the underlying generation quality, availability, and economics remain tied to third-party models. Its early public footprint also means prospective customers and investors should look for further evidence of retention, successful production usage, and reliability over time.
Note: Information based on publicly available sources at the time of writing, and summarized by AI.
Pitch your Startup, App or Hardware or post a Startup Event or Startup Job
About the Author
