LitLit AI Academic Workstation

What is LitLit?

LitLit is an AI-based academic workstation operated by Hunan Equator Galaxy Technology Co., Ltd.; it offers researchers, students, and institutions services for paper searching, analysis, review, writing, and knowledge management. According to the current product description, it has access to around 300 million papers from around the world.

Its core idea is to allow users to describe research questions in natural language, after which the search results are provided for further study, through literature maps, reviews, and knowledge bases. It is a tool to assist in research, but it cannot replace research design, peer review, or academic responsibility.

Current product portfolio

Natural language paper retrieval

Users do not need to enter only short keywords; they can directly describe the research topic, the target journal, authors, year, full-text requirements, or citation relationships. The system converts these criteria into tasks for filtering papers and displays the number of matches as well as their relevance level.

• Search for relevant papers by topic and full text content.
• Restrictions on periodicals, years, and other bibliographic criteria.
• Search for works by a specified author and associated collaborations.
• Track the references of a particular paper.
• Search for subsequent studies that cite the specified paper.
• Conduct in-depth searches around complex issues.

Natural language retrieval reduces the difficulty of crafting search queries, but it does not guarantee that all relevant studies will be identified. Those identified as having a \"perfect match\" or being \"highly relevant\" still require manual evaluation taking into account the title, abstract, full text, and research design.

How can search results be used?

The search results are not presented as a simple list; users can proceed to translate them, cite them, study them in detail, export them, or add them to a knowledge base. This approach is suitable for first creating a pool of candidate documents before moving on to in-depth reading and the writing of reviews.

1. Enter the research question and choose between quick or in-depth search.
2. Check the title, author, year, and journal of the matching paper.
3. Irrelevant literature was excluded based on the abstract and research methods.
4. Check citation and cited relationships to identify key studies.
5. Add important documents to the knowledge base or document map.
6. Export in the desired format and verify it in a reference management tool.

Quick search and in-depth search

The interface clearly indicates that deep searches are subject to daily limits or limits on the number of searches permitted per member; stopping a task in progress can also consume one of these quota units. The fixed number of searches cannot be understood in isolation from the specific account and subscription plan.

Literature Review and Academic Writing

The ability to conduct a literature review involves searching for and organizing papers related to a specific research topic, in order to create a review or draft paper that includes citations. The product regards “accurate citations” as an important feature, but users still need to verify that each citation indeed supports the corresponding argument.

• Identify the directions for literature search based on the research question.
• Select references from the candidate literature.
• Organize the research context, differing viewpoints, and evidence.
• Generate an outline structure and draft paragraphs.
• Organize citations and references within the main text.
• Save the results to my creation to continue editing.

Number of references and writing style

A large number of studies does not necessarily mean that a review is of high quality; what matters are the inclusion criteria, the quality of the methods used, and the consistency of the evidence. The systematically generated classifications, identified research gaps, and conclusions all require the researcher to read the original articles again.

Study of AI papers

For studying papers, it is possible to upload PDF files, and reading as well as Q&A sessions can be carried out based on the content of those papers. The upload page currently accepts only PDF files, with a maximum file size of 20MB.

1. Select or drag in the PDF papers that have permission to be processed.
2. Wait for the file upload and parsing to be completed.
3. Check whether the titles, sections, formulas, and charts are recognized correctly.
4. Ask questions regarding the methods, data, conclusions, and limitations.
5. Generate an interpretation report and return to the original text for verification.
6. Download in the desired format or control the sharing scope.

Scans, complex two-column layouts, formulas, annotations, and low-quality charts can affect the accuracy of interpretation. Papers related to medicine, law, and engineering safety cannot rely solely on summary-based responses.

Interpretation of reports and export

Literature map

A literature map starts with key documents, extracts citation relationships, creates relationship matrices, identifies core documents, and detects thematic clusters. It is suitable for examining the development of research, rather than simply arranging papers based on keywords.

• Establish a network using core papers as seeds.
• Distinguish between classic literature and derivative literature.
• View references and subsequent citations.
• Identify core papers with high connectivity.
• Identify theme groups and research branches.
• Export documents or add them to your personal knowledge base.

The number of citations is influenced by database coverage, name ambiguity, and the merging of documents; it cannot therefore serve as a direct indicator of research quality. Interdisciplinary topics and emerging fields may also be underestimated due to a lack of citations.

Personal knowledge base

The knowledge base is used to store papers, attachments, and research materials in folders, and it also handles tasks such as searching, displaying maps, and presenting analysis results. The platform offers functions like batch import, bookmarking, renaming, moving items, previewing, and analyzing data.

1. Create folders for projects, topics, or papers.
2. Collect papers from search results or import materials in bulk.
3. Upload the necessary attachments and give them consistent names.
4. Analyze the paper and check the metadata.
5. Link the research, reviews, and map results to the project.
6. Regularly export important results and keep local backups.

The storage capacity of the knowledge base is governed by the quota set for each plan; the fixed capacity allocated to each member is not listed on the public pages. It is necessary to export the data that still needs to be retained before the service is terminated, before a payment is overdue, or before the account is canceled.

Scientific Research Drawing Beta

Scientific plotting is currently marked as Beta on the main interface, indicating that the feature is still in the testing phase. Users can use it to explore graphical representations, but they should not assume that all graph types, export formats, and data processing methods are yet stable.

In scientific illustrations, the data range, units, errors, and statistical details must be retained; the conclusions should not be altered for aesthetic reasons. Before submitting, it is necessary to check the journal’s requirements regarding size, color space, font style, and accessibility.

LitClaw and Lili S1

LitClaw is marked as Beta in the sidebar of the workstation, while Lili S1 is an independent AI scientist product that is available for a limited beta test. Neither of these should be confused with regular paper searches or membership benefits.

According to the Lili S1 page, it allows users to take control of the research process, schedule computing resources, and run tasks for up to 100 hours; in addition, it offers more than 300 different workflows and capabilities. This describes the features available during the public beta phase, and it does not mean that all registered users already have access to them.

The complete process from topic selection to a review

1. Break down the research question into objects, variables, scenarios, and time range.
2. First, use a quick search to verify the keywords and candidate documents.
3. Run deep search after the range stabilizes.
4. Set inclusion, exclusion, and quality assessment criteria.
5. Add key documents to the knowledge base and document map.
6. Study each article in detail, examining its methods, data, conclusions, and limitations.
7. Select appropriate literature to establish the structure of the review.
8. Check each reference, number, and the relationship with the original text.
9. Disclose the use of AI in accordance with the requirements of the institution or journal.

Members, writing packages, and institutional plans

The current public page does not display fixed prices, automatic renewal options, refund policies, trial periods, or exact amounts; access to the price information also requires login. Even if the member name remains the same, previous promotional prices may no longer be valid, and the actual pricing shown on the settlement page and in the contract should prevail at the time of purchase.

Registration, login, and institutional accounts

Individual users can log in using a phone number verification code, account credentials, or the WeChat entry; some browsing and searching functions are available without registration. Full saving, uploading, as well as member and task features usually require an account.

An institutional plan allows for the configuration of team members, access rights, service settings, and institutional space data. Administrators have the authority to add or remove members, and the institution must specify in advance the details related to member data and account boundaries.

Data collection and model training

The platform will process prompts, search terms, task descriptions, conversations, papers, experimental data, code, images, tables, execution records, intermediate results, and final outcomes based on the functions being used. Users should remove any unnecessary personal information before uploading.

Without separate notification and proper authorization in accordance with the law, the platform will not use the original inputs, uploaded materials, or results of tasks for model training other than those related to the services requested by the user. The user’s rights to their business data are not transferred as a result of uploading, nor do unpublished research materials become public just because they are uploaded.

Third-party computing power and data locations

The cloud computing power, model inference, and code execution related to the Lili S1 system may involve third-party service providers in handling the necessary information. The platform states that all related cloud computing services, model services, servers, as well as activities related to the processing of personal data are carried out within China, with no cross-border data transfer taking place.

• The trustee may handle task instructions, documents, code, and runtime parameters.
• It may also handle task identifiers, intermediate results, outputs, and necessary logs.
• The platform requires that data be provided only to the minimum extent necessary.
• The trustee shall not use it for its own business or for unauthorized model training.
• Upon completion of the task, it is returned, deleted, or anonymized as agreed.
• For sensitive items, it is still necessary to consult the specific list of third-party services.

Data saving and deletion

Personal information is, in principle, stored for the shortest period necessary to achieve the intended purpose; web logs and transaction records are also subject to legal time limits. The data related to S1, as well as task records, intermediate results, and final outcomes, are retained for as long as it is necessary for the user to save them manually, to continue using the tasks, or to utilize the corresponding functions.

After the service expires, is terminated early, or due to unpaid fees, the user’s business data will only be stored for the period specified by the applicable rules; thereafter it may be deleted and cannot be recovered. Before canceling the account, it is necessary to download, export, or back up chats, favorites, search results, tasks, uploaded materials, and any generated content.

Intellectual property and the use of outcomes

Users may utilize the platform services and the results generated by them for personal research, teaching, institutional scientific research, corporate R&D, and other legitimate purposes. The rights to the platform, models, algorithms, software, interfaces, templates, general capabilities, and existing materials remain with LiLi or the respective owners of those rights.

• Uploading does not change the original ownership of rights to the papers, data, and code.
• Users must have the right to allow the platform to process the uploaded materials.
• The results can be used in accordance with the agreements and the specific rules applicable to the products.
• The generated content is not guaranteed to possess originality under copyright law.
• The generated content is also not guaranteed to be completely different from other people’s content.
• The platform services themselves cannot be resold, packaged, or licensed to third parties.
• Before sharing publicly, it is necessary to verify third-party copyright and confidentiality obligations.

Academic Integrity and AI Attribution

Users are not allowed to fabricate or alter research data, create fake documents and citations, engage in plagiarism, or attempt to bypass ethical review requirements, rules regarding the use of AI, and submission guidelines. The authors of a paper remain fully responsible for the research methods, conclusions, authorship, and citations used.

The platform will add explicit or implicit identifiers to AI-generated synthetic content in accordance with legal requirements; users are not allowed to delete, alter, or conceal such identifiers improperly. When releasing this content to the public, it is also necessary to comply with the labeling requirements set by the respective institution, journal, and applicable regulations.

Privacy and security

The privacy policy states that the platform employs measures such as encrypted transmission, isolation, anonymization, data classification, access control, and employee confidentiality. It also makes it clear that the Internet environment is not absolutely secure.

• Account registration may collect phone numbers, passwords, and verification codes.
• Some services may require real-name verification in accordance with the law.
• The website processes information related to devices, IP addresses, logs, and click streams.
• Payment, customer service, and security functions will handle the corresponding records.
• Users under 14 years of age require the consent of a guardian.
• Users may legally request access, correction, deletion, and cancellation.

API, SDK, and open-source status

Supported platforms and languages

The interface is primarily designed for Chinese users, while the search results and the documents themselves may be in multiple languages. The language used in the interface, the language used for searching, the language used for parsing PDFs, and the language used for generating content are all different aspects; it is necessary to test each of them using actual papers.

Which users are it suitable for

• Graduate students who need to quickly establish a list of potential papers.
• Students preparing a literature review and a project proposal.
• Researchers who track citation networks and research trajectories.
• Teams that need to study papers in bulk and generate reports.
• Universities and research institutions that are developing institutional knowledge bases.
• Personnel responsible for corporate R&D and technical intelligence work.
• Long-distance research teams that have qualified for the S1 beta test.

Main advantages

• Natural language retrieval can combine various paper criteria.
• Search, study, review, maps, and knowledge bases form a closed loop.
• The paper review report supports various file formats.
• Literature maps help understand citation and thematic relationships.
• The user business data and the boundaries for model training are defined quite clearly.
• Organizational teams and autonomous research agents provide pathways for expansion.

Capacity boundaries

• Database coverage does not guarantee that all papers are included.
• Matching tags cannot replace the manual inclusion criteria.
• The generated content may contain errors in facts, citations, and data.
• PDF parsing is affected by the file quality and the 20MB limit.
• Deep search, writing, and storage are subject to limits set by the plan tier.
• The fixed member prices and refund rules are not made publicly available.
• Scientific plotting tools, LitClaw, and S1 are still in the testing or public beta phase.
• The public APIs, SDKs, and open-source licenses have not been confirmed.

Frequently Asked Questions

Can LitLit be used for free?

The public entry point allows access to some functions, but the specific number of free uses, as well as the options for advanced searching, in-depth analysis, and storage capacity, are indicated based on the account details. The basic version should not be considered to offer all functions indefinitely at no cost.

Can Lili generate a literature review with citations?

It is possible to create a reviewed article or a draft paper with citations based on the searched literature. However, it is still necessary to verify each source, page numbers, data, and the relationships between various elements by referring to the original texts.

What types of files are supported for paper research?

The current upload page supports PDF files, with a maximum size of 20MB per file. The compatibility of other document formats and scanned images needs to be tested within the account.

How many references are needed to generate a paper?

The current rules require a minimum of 30 and a maximum of 150 papers; the quick mode can handle up to 30 papers. Generating papers also requires an available quota for writing packages.

Will the council use uploaded papers to train the model?

Without separate explicit authorization, the original prompt words, uploaded materials, as well as unpublished research data or results will not be used for training general models or base models intended for other users.

Is the Lili S1 the same as the AI academic workstation?

They are different. S1 is designed for long-term research processes and computing sandboxes, and it is currently available in a limited beta version; workstations, on the other hand, offer functions for daily searching, studying, reviewing information, and managing knowledge.

Does Lili provide public APIs?

The user agreement covers API services, but at present there are no public documents outlining the procedures for individuals to apply for these services, nor information regarding keys, pricing, or rate limits.

Is Lili an open-source project?

At present, the official product code, model weights, or open-source license have not been confirmed. The web services, APIs, or testing functions do not mean that the product itself is open source.

How to add a FAQ Schema

This page already provides FAQs that are visible to users; the site template can use these questions and answers to generate structured FAQ data. Script tags are not inserted directly into the text, in order to meet the security requirements for content fields.

1. Select only the actual, visible questions and answers on the page.
2. The page object is set to FAQPage.
3. Each question is set as Question.
4. The answer is written to acceptedAnswer and set as Answer.
5. The limits, file restrictions, and data rules must be consistent with the main text.
6. The visible FAQ is updated in real time when there are changes in the member status or beta testing status.
7. After publishing, check for grammar errors, repetitions, and search engine guidelines.

Summary

The value of LitLit lies in its ability to integrate natural language paper searching, PDF analysis, literature reviews, citation mapping, and personal knowledge bases into a cohesive workflow. It is suitable for individuals and institutions that need to work with large amounts of academic literature.

In actual use, retrieval strategies, verification of the original text, and academic standards should be the guiding principles; AI is only responsible for speeding up the process of organizing and generating content. Before purchasing a membership, it is necessary to check in the account details the options related to advanced searching, writing tools, storage space, prices, and refund policies.