Blue light fish

What is a blue light fish?

BlueLightFish is an AI-assisted intellectual property platform operated by Beijing BlueLightFish Intelligent Technology Co., Ltd.; it is designed for patent agents, intellectual property firms, corporate legal teams, and technology transfer teams.

The platform assigns tasks such as analyzing repetitive materials, comparing technical features, creating drafts, and organizing search results to AI, but insists that professional judgment, legal strategies, and the final approval should be left to specialists.

Main products and applicable tasks

OA Elf: Response to Patent Examination Comments

Review of document parsing

• The user imports the examination opinions, the current application documents, and the referenced comparative documents; the system then extracts the examination conclusions, relevant legal provisions, points of dispute, and key paragraphs.
• Structured results are used to quickly establish the overview of a case, but scan quality, file integrity, and complex tables can affect the parsing process.

Comparison of technical features

• The system compares the technical features in the claims with those in the reference documents, to assist in identifying the distinguishing features and the corresponding evidence thereof in the description.
• The comparison results should be verified item by item by the agent against the original documents; content that was not detected by the model cannot be considered non-existent.

Argumentation strategies and draft responses

• The OA Elf can organize potential arguments around criteria such as novelty and creativity, and generate draft responses that can be further revised.
• The platform emphasizes that the entire output process can be modified and traced, making it suitable for focusing attention on evidence, legal principles, the scope of modifications, and risk assessment.
• The arguments presented by the system are not guaranteed to be accepted by the reviewers, nor do they guarantee a approval rate.

OA response workflow

1. Create a case and verify the application number, version of the application documents, deadline for responses, and type of review comments.
2. Upload the review comments, the application text, and all supporting documents, to confirm that the page numbers and content have been correctly identified by the system.
3. Review the key points of the review and the corresponding technical features, and add any terms, structural relationships, and context that are missing in the system.
4. Identify the distinguishable features that can be used for filtering, and check whether the instructions provide clear and direct reasons for the modifications.
5. Let the system generate draft arguments and responses, after which the attorney adjusts the claims and the supporting arguments.
6. Check for out-of-range modifications, logical contradictions, term consistency, accurate citations, and deadlines.
7. It is reviewed and finalized by a licensed patent agent, before being submitted through official application channels.

XinZhiWrite: Draft of new patent application

Understanding of technical instructions and breakdown of the plan

• A technical briefing on core intelligence writing and reading techniques, covering identification of components, steps, data flows, causal relationships, technical outcomes, and potential innovations.
• Users should clarify the core innovations, essential technical features, optional implementation methods, as well as the technical boundaries that should not be expanded or omitted.
• The more complete the information provided and the clearer the terminology, the easier it is for a system to take on a usable structure; in complex professional fields, technical expertise is still required.

Search and creative analysis

• The platform can combine multiple documents to generate clues for search and analysis, and assist in determining the differences between a proposed solution and existing technologies.
• The search results are influenced by database coverage, keywords, classification numbers, language, publication date, and algorithms; they cannot replace a thorough novelty check or legal assessment.

Writing and interactive editing

• The system can generate a draft of the claims framework and the description, allowing users to make direct modifications on the page and to further refine them through interactive processes.
• The output can serve as a starting point for working as an agent; applications should not be submitted directly without verification.
• Technical features, experimental results, or examples that lack material support should be removed or supplemented, in order to avoid the risk of submitting dishonest applications.

Patent drafting workflow

1. Organize the complete technical specifications, drawings, experimental data, alternative solutions, and existing technologies.
2. Create a case and upload the relevant materials, specifying the type of application, the technical field, and the core solution that wishes to be protected.
3. Review the technical breakdown of the system, and correct terms, component relationships, step sequences, and innovative aspects.
4. Based on the search results, determine the essential features of independent claims as well as their hierarchical relationships.
5. Draft the claims and description, and add implementation examples through dialogue or an editor.
6. Check the support relationships, clarity, uniqueness, reference numbers, consistency of terminology, and the degree of completeness of the disclosure.
7. It is reviewed jointly by technical personnel and certified agents; once confirmed, the formal application process is initiated.

One-stop trademark services

The scope of trademark searches and their latest update dates can be found in the list corresponding to the respective country or region; the results displayed on the page are for reference only and do not have legal validity.

Patent translation

• Machine translation makes use of AI engines designed for patent texts, supports multi-language processing, and can be adapted to various deployment methods.
• Human translation is provided by a professional team; the scope of work, delivery time, formatting, terminology database, and quality standards are determined based on the specifics of each project.
• When translating patents, it is essential to pay close attention to the terminology used in the claims, negative expressions, numerical units, reference numbers in the drawings, and the consistency with related documents.
• The percentage reduction in costs indicated on the promotional page is based on the specifications of the product; the actual costs and quality depend on the language used, the length of the text, the field in question, and the level of scrutiny applied.

Price and purchase methods

The public page does not display the unified subscription price, case quota, renewal options, refund policies, or enterprise deployment costs; to access the ordering page, one must first log in.

Before making a purchase, it is necessary to verify on the settlement page or in the contract the price including taxes, the duration of the service, the available modules, the number of projects that can be carried out, any additional fees, the conditions for refunds, and the delivery standards.

Platform and deployment methods

Precautions for private deployment

• The platform supports pure local deployment; the models can operate independently within the organization’s intranet without the need to connect to the external network.
• When making purchases, it is necessary to ensure that the models, knowledge bases, logs, backups, administrator privileges, and monitoring data remain within the designated environment.
• It is also necessary to clarify the requirements for servers and graphics cards, the level of concurrency, the installation time required, the methods of upgrading, procedures for handling failures, the service level, and what happens to the data when the service is discontinued.
• “The principle of data staying within the domain” should be based on the final technical architecture, network testing, and contractual agreements as criteria for acceptance.

API, SDK, and open-source status

• At present, there are no confirmed API documents available for the public, nor developer keys, SDKs, pricing information, or service levels.
• Private deployment and enterprise integration do not imply the provision of public APIs; the scope of these interfaces must be confirmed separately with the sales and technical teams.
• To date, it has not been confirmed that the Blue Light Fish platform uses an open-source license, and no official code repository proving that its core components are open source has been found.
• Even if examples, plugins, or separate repositories appear later on, this does not imply that the online platform, models, or trademark data as a whole are open source.

Privacy and data security

Collected information

• Registration may involve handling phone numbers, account details, passwords, and third-party login information, while purchases involve processing order details, payment information, and invoices.
• Real-name verification may involve information such as name, identity documents, business licenses, bank accounts, or facial recognition data, depending on the verification method chosen by the user.
• Websites and mini-programs may handle device information, network logs, search queries, access times, permissions, and operation records.
• Technical instructions, application documents, trademark designs, and case-related materials are considered highly sensitive business data; before uploading them, it is necessary to verify the relevant authorization from the organization as well as any confidentiality requirements.

Business data and storage

• The privacy policy states that users have control over the business data they upload, generate, process, and store, with the platform handling it according to the users’ instructions.
• The product FAQ states that the technical details provided are intended solely for the analysis and preparation of current cases; they are not stored in any public database nor used for model training.
• The product’s FAQ also states that the original documents are automatically deleted two months after the case has been resolved; other personal information is retained for as long as it is necessary to achieve the intended purpose.
• Different files, drafts, logs, backups, and those required for legal retention may have varying retention periods; institutional users should confirm each of these aspects in the contract.

User rights and security measures

• Users can view, correct, and delete certain information in their personal account, revoke authorizations, and request account cancellation.
• Upon account cancellation, the platform states that personal information will be deleted or anonymized, unless required by applicable laws and regulations.
• The policy outlines measures such as encryption, access control, security auditing, and event notification, while emphasizing that internet systems cannot guarantee 100% security.
• In principle, the information collected within China is stored within the country; if certain services require data to be transferred outside the country, this must be communicated in advance and the corresponding procedures must be followed.

Content creation, copyright, and commercial use

• Users must ensure that the technologies, patterns, evidence, and documents they upload are handled in accordance with legal requirements, and they are responsible for the content of the business data.
• The agreement prohibits the copying or commercial use of the platform’s materials without written permission, and it also forbids access through crawlers, robots, and other automated methods.
• The agreement states that the user remains the owner of the content they publish, but grants the platform the right to display, use, and distribute it within the scope of the service.
• The platform applies explicit and implicit markings to AI-generated content, and users are not allowed to remove, conceal, alter such markings or circumvent them in any unlawful manner.
• The generation of an identifier does not imply that the platform guarantees the authenticity, accuracy, ownership, or suitability of the content.
• Whether the deliverables provided by the customer can be used for commercial purposes, distributed to third parties, and who holds the rights to them shall be determined in accordance with the specific order and contractual agreements.

Which users are it suitable for

• Licensed patent agents who need to improve the efficiency of drafting new patent applications and responding to OA requests.
• Intellectual property institutions that need to standardize case workflows, quality control, and knowledge management.
• A trademark team that handles tasks such as trademark searches, monitoring, reviews of rejected applications, and handling objections.
• A foreign-related team that requires multilingual patent translation and manual review.
• Enterprises, universities, and research institutions that have strict requirements regarding confidentiality when it comes to technical details and prefer to deploy systems on an intranet.

Product advantages

• Patent drafting, responses to examination requests, as well as trademark and translation services all fall within the same intellectual property service framework.
• The output can be further edited and adjusted interactively, making it suitable as a structured draft for professionals.
• It supports the combined analysis of multiple files, thereby reducing the need for manual sorting and repeated comparisons.
• It offers a purely local, private deployment option, suitable for organizations that have requirements regarding internal networks and data boundaries.
• The platform clearly states that the generated drafts must be reviewed by professionals, which helps to define the boundaries of responsibilities between humans and machines.

Capacity limits and usage restrictions

• AI may miss certain features, misinterpret causal relationships, contain logical flaws, or produce statements that have no factual basis.
• Incomplete technical instructions, inconsistent terminology, or poor scanning quality can significantly affect the results.
• There are time differences in coverage and updates between trademark and patent databases, and the search results do not have legal validity.
• The output cannot replace the confirmation provided by licensed agents, technical personnel, and applicants, nor does it guarantee approval or a successful review.
• The efficiency and cost-saving percentages shown on the public pages are part of the product’s promotional information, and should not be regarded as guarantees that all users will be able to achieve.
• There is no unified and public information available regarding prices, quotas, refunds, and service levels; written confirmation is required before making a purchase.

Frequently Asked Questions

What is the blue light fish used for?

It is an AI-powered platform designed for professionals in the field of intellectual property, offering services such as draft patent preparation, responses to official requests, trademark searches and related documents, patent translation, and private deployment options.

Can blue light fish file a patent application directly?

The AI-generated draft cannot be used as the final application document; the platform requires that it be reviewed, modified, and finalized by a licensed patent attorney before it can proceed to the official application process.

Can the OA Elf guarantee a successful response?

No, it is used to analyze the review comments, compare various features, and generate draft responses; the final outcome still depends on the facts of the case, legal judgments, the relevant evidence, and the decisions made during the review process.

Will the uploaded technical documentation be used to train the model?

According to the platform’s current guidelines, the statement of requirements is intended for use only in the specific case at hand; it is not inserted into public databases nor used for model training. For highly sensitive projects, it is still necessary to define data boundaries and confidentiality responsibilities through a contract.

How long is the original technical briefing document kept?

According to the product FAQs, the original documents are automatically deleted two months after the case has been resolved; however, it remains to be determined whether the same time frame applies to the generated versions, logs, backups, and those documents required for legal purposes.

Does the Blue Light Fish support local deployment?

It supports pure local, on-premises deployment for organizations; the specific hardware requirements, interfaces, concurrency limits, maintenance needs, upgrade procedures, and acceptance criteria must be determined as part of the customized solution for each enterprise.

Can the results of a trademark search be used as a legal conclusion?

No, the page clearly states that there may be a delay in updating the data; therefore, the search results are for reference only and do not have any legal validity.

How much is the blue light fish?

The public page does not display the unified current prices and quotas; to view information on online products, it is necessary to log in. For private deployments and customized services, you should contact sales for a written quote.

Is there a public API or an open-source version for the Blue Light Fish?

At present, there are no confirmed open APIs, SDKs, or open-source licenses for the core products; moreover, enterprise integration or private delivery does not equate to public development interfaces or open-source products.

How to add a FAQ Schema

On this page, common questions are displayed using hierarchical headings, with the answers visible right next to the corresponding paragraphs; the website can use this information to generate structured FAQ data on the server side.

Structured data must be consistent with the content visible on the page; it should not include hidden prices, approval rates, legal conclusions, links, or unverified APIs and commercial promises.

Summary

The Blue Light Fish is suitable for converting repetitive tasks in patent drafting, OA responses, trademark handling, and translation into auditable AI-assisted processes.

Its proper role is to serve as an efficient tool for professionals; any application, response, evidence, deadline, or decision related to commercial use requires manual verification and ultimate responsibility.