Yinsi AIGC

What is Insee AIGC at the moment?

In its early stages, Insee AIGC was known as InsightGPT; the current brand name is Insee AI. The old entry point now displays the new Insee AI marketing platform, directing users to DynSight and InFlow.

YinSai AI is part of the AI marketing system owned by Guangdong YinSai Brand Marketing Group Co., Ltd.; DianShang and YingLiu are developed and operated by Guangdong Creative Hot Store Internet Technology Co., Ltd., a wholly-owned subsidiary of the group. It is not a simple chatbot, but rather a multi-modal content platform designed for e-commerce businesses, brands, advertisers, and creators.

Changes in brands and products

The aforementioned products are all part of InSai Group’s AI strategy, but they differ in terms of access points, users, permissions, and procurement methods. Registering as a business partner or streamer does not automatically entitle one to InClaw, ZhiMou AI, or InSai AI marketing services.

DynSight e-commerce AI workspace

E-commerce templates and workflows

• The home page of Dianshang offers a template library; the currently available categories include clothing, jewelry, sets of similar images, reproductions of images popular among internet influencers, reproductions of best-selling products, e-commerce product detail pages, and main product images for e-commerce use.
• Users can use the “Start in One Click” template to organize product images, reference visuals, copywriting, and size requirements into a standardized workflow.
• The output includes product images, details page materials, promotional content, product videos, and marketing videos; the specific format and quantity depend on the workflow chosen.
• Templates help to standardize the structure and procedures, but they do not automatically ensure platform approval, ad compliance, or sales conversion.

Community and reuse

• Explore the community to view selected artworks, as well as content organized into categories such as e-commerce, film and television, advertising, animation, short narrative films, MVs, new media, and tutorials.
• Users can access “Start with One Click” from publicly available works, utilizing the canvas structure and creation process to reduce the time required to set up a new project from scratch.
• When replicating templates or reference works, it is still necessary to check the licensing status of the reference images, products, characters, fonts, and music; just because something is visible to the public does not mean it can be used commercially.

InFlow’s creative ecosystem

• Yingliu is aimed at AI creators and brands; it brings together popular AI models for the creation of graphic designs, videos, short dramas, and other creative content.
• Creators can use models and workflows to produce content, while brands can release this content or meet various commercial needs, thus establishing a connection between content supply and marketing requirements.
• The group stated that Yingliu also plans to integrate influencer marketing with intelligent advertising strategies, in order to create a more seamless process for content creation, distribution, and conversion of results.
• Insei AI marketing services are provided and developed by a team of creative experts through a collaborative approach between humans and machines; they represent enterprise-level membership or customized services, and cannot be considered equivalent to ordinary self-generated solutions.

Current marketing content scenarios

Current model and new capabilities

• The current brand page displays information on various models or versions such as MiniMax H3, SeeDance 2.0, SeeDance 2.0 mini, SeeDance 2.5, Vidu Q3, and Suno V5.5.
• Camera presets are used to apply fixed camera movements during production, thereby reducing the effort required for users to repeatedly describe the camera angles in a film.
• Pose transfer applies the pose of a reference figure to the target content, and is suitable for testing consistency in clothing, e-commerce models, and characters.
• 720 panoramic imaging is used to create spatial views, which are suitable for scene previews; however, the actual dimensions, geometry, and product details still need to be verified.
• The model name, price, resolution, input limitations, and available regions will be updated; the appearance of certain elements on the page does not mean that all accounts have the same permissions.

Complete usage process

1. Choose between DianShang or YingLiu depending on the task: for product marketing, DianShang should be used first; for cross-category creative efforts and brand collaborations, YingLiu can be considered.
2. Browse public templates and community examples to determine the output type, reference materials, models, format, and estimated time required for completion.
3. Register and log in, and complete the identity verification required by the platform; account credentials are necessary for the core functions of creation, publishing, and interaction.
4. Create a new canvas or reuse a template, and upload authorized product, character, brand, as well as audio and video materials.
5. Fill in the brand brief, product features, target audience, distribution channels, style, and prohibited content, avoiding generic instructions that consist of just one sentence.
6. First, create a prototype using low-cost parameters to check the composition, text, accuracy of the people and products, as well as the content labels.
7. Through workflow iteration scripts, images, shots, and videos, model versions, parameters, milestones, and failed tasks are recorded.
8. Complete the review of copyrights, advertisements, portraits, trademarks, platform guidelines, and AI identifiers before downloading or publishing.
9. For team projects, the original materials, authorization documents, records of manual modifications, and the final approved version should be preserved.

Prices, tiebreakers, and corporate solutions

The “minimum” price refers only to the lowest price displayed when certain specifications, resolution levels, and campaign conditions are met; it cannot serve as a uniform rate for all video tasks. Factors such as production duration, retry attempts in case of failures, audio quality, high-definition settings, concurrent processing, and downloading can affect the actual cost.

On the current page without a login, the complete details regarding conversion rates to RMB, the free usage quota, membership levels, and expiration dates are not available. Before making a payment, it is necessary to check the account page for information on how resources are consumed item by item, as well as details on orders, refund conditions, and team benefits.

Refund rules

• The service agreement stipulates that tie-breakers, membership packages, and other paid services are, in principle, non-refundable; it is necessary to confirm the products and models before making a payment.
• A refund can be negotiated only in cases where there are significant quality issues with the platform services; for example, if the core functions are unavailable for more than 24 hours continuously, or if there is a single interruption lasting more than 12 hours without any alternative services available.
• In the event of significant defects, the refund is determined on a proportional basis based on the amount of service that has not been used; it does not mean an automatic full refund.
• Content generated by the model that does not meet an individual’s aesthetic standards, is similar to other results, or contains flaws that can be corrected manually, cannot generally be considered a service disruption.
• In the event of an anomaly, save the task number, the model, parameters, transaction history, time, and error messages, and contact the platform’s customer service as soon as possible.

Copyright, Commercial Use, and AI Indicators

• The service agreement states that, after full payment, it is possible to use the images, videos, and text generated by the platform within the bounds of what is legal and compliant, provided that the restrictions regarding the licensing of those materials and templates are respected.
• Users only have rights to their own original contributions and to those outputs that are permitted by law; the platform’s materials, templates, fonts, code, and brand identifiers remain the property of the platform or their respective owners.
• The login page indicates that the development team retains the rights to the created works and suggests seeking detailed information regarding authorization; moreover, the scope of paid use as specified in the agreement needs to be clarified in relation to specific products.
• Before commercial use, it is necessary to keep the payment page, the licensing details for the materials, the terms related to the models, and the corporate contract; one should not rely solely on the fact that something is downloadable or watermark-free.
• References to images, characters, trademarks, products, music, and fonts must be authorized; the platform’s content review does not constitute a guarantee against infringement.
• The platform will add explicit and/or implicit identifiers to the AI-generated content; users are not allowed to alter, delete, or modify it without permission, and such content must be clearly labeled in accordance with the applicable rules when it is published.

Input and output content licensing

The agreement grants the platform broad rights to use the content entered, output, and published by users, including terms that are global in scope, permanent in duration, irrevocable, non-exclusive, and permit sub-licensing. The purposes of these rights include providing services, optimizing products, training models, and ensuring security.

• Trade secrets, unreleased advertisements, customer lists, internal prices, and materials covered by confidentiality agreements should not be uploaded directly.
• Works made publicly available may also be copied, translated, compiled, distributed, displayed, or adapted by the platform and its partner platforms, for the purpose of promoting the creative works.
• The privacy policy also states that the input data is used solely for generation purposes, and shall not be employed for any other purposes without consent; data that has been thoroughly de-identified so that it cannot be identified again, along with the content generated, can be used for model optimization.
• The scope of application of agreements and privacy policies varies; companies should use written contracts to clarify whether such data is to be used for training purposes, what the exit procedures are, and how customer data is isolated.

Privacy and data storage

• The core functions of creation and publishing require registration using a mobile phone number and verification of one’s real identity; third-party logins may handle aspects such as the profile picture, username, and mobile phone number related to authorization.
• The platform handles input text, uploaded images or audio/video files, operation commands, searching and browsing, publishing interactive content, as well as device, log, and transaction information.
• Face and voice materials are used solely for creating portrait photos, digital avatars, etc., and not for identifying individuals or verifying personal information.
• Personal information is stored within China; if cross-border transfer is necessary, the policies state that the required procedures will be followed and explicit consent will be obtained in advance.
• Phone numbers are retained until the account is canceled; posted content remains in place until the user deletes it or cancels the account; service logs and device information are kept for 6 months after the service ends.
• Transaction information must be retained for at least 3 years from the date the transaction is completed; other information that has exceeded this time limit is, in principle, deleted or anonymized, unless otherwise provided by law.
• Users can request to view, copy, correct, delete their data, withdraw their consent, and close their accounts; once an account is closed, the relevant data is deleted or anonymized immediately, in principle.

Platforms and integration

API, GitHub, and open-source status

• Currently, no documentation available for the public regarding API’s, SDKs, key applications, and usage instructions for InSai AI, DianShang, or YingLiu developers.
• The platform’s aggregation of third-party models merely means that users can invoke them through the interface; it does not imply the resale of those APIs to developers.
• No official code repository has been found that demonstrates the full open-source nature of the InSai AI platform, the workflow codes, or the marketing models developed in-house; therefore, it should be considered a proprietary product.
• The name of a model appearing on the functionality page does not indicate that its weight, training code, or commercial license are made available by InSight; it is necessary to comply with the terms set forth by the provider of that model.

Suitable for users and scenarios

• E-commerce sellers: Create bulk quantities of main images, product detail pages, promotional graphics, product videos, and marketing materials.
• Brand marketing team: Creates a collaborative production process that covers briefs, scripts, visuals, and advertising videos.
• New media management: Creating voiceovers, creative short videos, and advertising materials, while rapidly developing multiple versions of them.
• AI creators: Use multiple models and workflows in Yingliu to produce graphics, short dramas, animations, and MVs, while also exploring brand requirements.
• Games and Entertainment Team: Creates character designs, promotional materials, derivative images, and concept videos; conducts IP audits prior to the official release.

Advantages and limitations

Main advantages

• It has evolved from the early universal marketing approach to a model based on specialization in sales and content creation, resulting in a clearer structure for e-commerce workflows and various types of creative tasks.
• By combining templates, canvas communities, and multiple models, users do not need to start building each project from scratch with empty parameters.
• Insa Group’s marketing expertise, human-machine collaborative services, and brand creator ecosystem can provide enterprises with support beyond self-service tools.

Current restrictions

• The full price, free match points, model usage, and team benefits require login or a custom request; there is limited transparency.
• The quality, speed, resolution, and price of different models change rapidly, which makes previous tutorials easily become outdated.
• The agreement grants the platform broad rights to use the input and output data; however, trade secrets and customer materials require additional compliance checks.
• The login prompts and the terms of service regarding the rights to the created works need to be interpreted in light of the specific authorizations granted; it is not appropriate to assume anything regarding commercial use.
• The public API, SDKs, and source code are not available, so they are not suitable for use as developer model interfaces.

Frequently Asked Questions

Are Insee AIGC and InsightGPT the same product?

InsightGPT was the original name; the current brand name is now YinSai AI. The old entry point now displays the new YinSai AI page, which is connected to DynSight and InFlow.

What is the difference between Dian Shang and Ying Liu?

Dianshang focuses on product images, videos, and e-commerce workflows, while Yingliu is dedicated to the creation of graphics, videos, short dramas, and the ecosystem of brand creators. Both can use similar models, but their tasks and platform structures differ.

Can it be used for free?

Some pages and basic functions can be accessed anonymously, and the brand pages advertise a cost-free experience; however, the specific free features, functions, and time periods are not publicly specified. To use the core functionalities, it is necessary to register with a real name.

How much is the minimum cost for video generation?

The current offer shows that SeeDance 2.0 in 720p resolution is available at just 0.49 yuan per second. This price is subject to constraints related to the model, video quality, as well as the conditions of various activities and tasks; it therefore does not apply to all video generation scenarios.

Can match points and membership fees be refunded?

In principle, no refunds are given. A refund for the unused portion can only be negotiated in cases where the core services meet the criteria for a significant disruption as defined in the agreement, and no alternative services are available.

Can the generated content be used for commercial purposes?

The payment-based agreement permits use within the bounds of legality, compliance, and authorized terms; however, the template materials and the login page contain additional information regarding rights. Before using them for commercial purposes, it is necessary to verify the specific licensing terms for each product and keep written proof of that.

Will the uploaded content be used for model training?

The agreement grants the platform extensive permissions for product optimization and model training; the privacy policy also allows the use of strictly de-identified data that cannot be re-identified, as well as generated content, for model optimization. Sensitive customer data must first be subject to a written agreement specifying its exclusion or isolation.

Are APIs or open-source code provided?

At present, no public APIs, SDKs, or official open-source repositories have been found. Aggregating third-party models on a website does not equate to providing open development interfaces or source code.

Summary

Yinsi AIGC has evolved from InsightGPT to the Yinsi AI marketing system; it is now more suitable for creating e-commerce content using DingShang, or for carrying out flat-design and video creation with multiple models on YingLiu. Before using it, one should first check the current pricing for different models, and then pay close attention to the rules regarding refunds, input/output permissions, the commercial use of materials, and the requirements related to AI identifiers.