Unriddle

What is Unriddle?

Unriddle was an AI research assistant designed for scientific research, learning, and professional investigations; it has now been renamed Anara. The entry point for the old brand leads to Anara, and its original functions such as paper reading, PDF Q&A, literature organization, and citation-based writing have been expanded to form a more comprehensive research workspace.

Anara allows users to upload papers, web pages, videos, and other materials; it also enables searching in sources such as PubMed, arXiv, JSTOR, etc., facilitating searches between personal databases and external information. The system’s responses indicate the sources of the evidence, helping users to go back to the original texts for verification, rather than providing only conclusions that cannot be traced back to their sources.

What is the relationship between Unriddle and Anara?

Anara is the new brand name for Unriddle; it is not a separate tool. Existing Unriddle users should refer to their current Anara account, the help center, and the pricing page.

The tool directory retains the Unriddle name in order to accommodate existing users and past searches, but the functions and packages listed there are based on the current Anara standards.

Upload papers and read PDFs

Users can import research materials such as PDFs, and use the built-in reader to highlight text, add annotations, pose questions, and locate specific passages. The AI can explain terms, summarize chapters, identify research methods, and extract results and limitations, making it suitable for initial reading of papers and for screening literature.

Scanned PDF files require OCR in order to be interpreted reliably; complex tables, formulas, images, and multi-page structures may still lead to errors in interpretation. Before citing any numbers or conclusions, it is necessary to open the original document and verify the page numbers, table headings, and context.

Multi-document Q&A

By placing multiple papers in the same folder or conversation, Anara enables cross-document comparison of methods, samples, conclusions, and discrepancies. It is suitable for answering comprehensive questions such as “Which studies used the same dataset?” and “Why do different papers arrive at opposite conclusions?”

The integration of multiple documents is influenced by the upload scope and search results; paragraphs that are not retrieved will not automatically be included in the answer. For systematic reviews, it is still necessary to keep the complete search queries, database scope, inclusion/exclusion criteria, and filtering records.

Natural language search in databases

Ask AI allows users to search through personal databases using natural language, without the need to remember file names or exact keywords. Users can ask where a certain concept appears, which paper uses a particular method, or retrieve relevant information from previously imported data.

Academic paper search

The Search Papers agent can locate papers in PubMed, arXiv, JSTOR, and other supported scientific databases. The search results can be added to a database, read, annotated, and cited, thereby reducing the need to switch between multiple pages.

Database coverage and full-text access rights vary depending on the source; finding a record does not mean that the full text can be obtained free of charge. For thorough research, it is also necessary to use professional databases recognized in the respective field for additional searches.

Websites and research searches

Research agents can make use of personal databases, academic databases, and the public internet simultaneously to carry out multiple searches and analyses in order to address complex issues. The Max solution also offers a Deep Search option, which enables more comprehensive parallel exploration of databases and web pages.

The quality of online information varies greatly. For policy, clinical, legal, and market-related data, it is essential to verify the original source; one should not rely solely on secondary summaries generated by AI.

AI response with citations

Anara emphasizes that each main statement can be traced back to its source. Users can view the corresponding files and excerpts to determine whether the AI has understood the evidence accurately.

Compared to ordinary chatbots, this design is more suitable for research and verification.

The fact that a citation is present does not necessarily mean that the reasoning is correct. AI might mistake correlation for causation, ignore constraints, or misstate the source; therefore, important details still need to be verified manually one by one.

AI notes and writing

The built-in note editor allows users to organize reading summaries, research questions, drafts of literature reviews, and project documents. In the older Unriddle version, entering specific shortcuts would enable text generation based on the context.

Currently, Anara emphasizes even further the need to include verifiable sources for each argument when writing.

Text generated by AI must be revised by the author, and it is necessary to comply with the disclosure policies set by schools, journals, and funding agencies. The system cannot replace original analysis, nor should it generate unverified research conclusions on its own.

Quote generation

After selecting text in the editor, it is possible to search the database for relevant evidence and generate citations. Before creating a citation, it is necessary to verify the author, title, year, journal, DOI, and citation style, to ensure that the source material indeed supports the sentence in question.

Knowledge graph

Knowledge graphs display database entries along with the connections identified by AI, helping users understand the relationships between authors, topics, concepts, and research. They are useful for exploring new fields and discovering connections that might have been overlooked; however, the relationships generated automatically are not equivalent to verified academic facts.

PDF annotations and tags

Users can highlight text and add annotations in PDF files, and manage documents using folders and tags. Tags can also be used to link files or folders within the editor, making them useful for creating long-term databases of documents, project-related materials, and team knowledge bases.

Connecting Zotero with Mendeley

Plans Plus and higher support the connection of Zotero and Mendeley, allowing existing reference libraries to be integrated directly into the Anara workflow and reducing the need for re-uploading. After importing, it is necessary to check that attachments, DOIs, duplicate entries, and the structure of collections are all complete.

Research database connector

Plus offers connections to PubMed, bioRxiv, medRxiv, ClinicalTrials.gov, PDB, ChEMBL, Open Targets, as well as Benchling and Synapse; it is particularly suitable for research in biomedicine, pharmacology, and life sciences.

Different connectors have varying data permissions, update frequencies, and field completeness; therefore, when using clinical or pharmaceutical information, it is necessary to consult authoritative databases for verification.

Connect cloud storage to collaboration tools

Pro adds connectors for Google Drive, Notion, OneDrive, Slack, Microsoft Teams, and Dropbox, allowing disparate research documents to be imported into a unified database. When connecting to corporate accounts, the least necessary permissions should be used, and the scope of shared folders should be clearly defined.

Selection of AI models

The free version has the model selected automatically by the system; versions Pro and above allow switching between various top-quality models.

The Pro models listed in the current help center include the GPT 5.6 series, GPT 5.5, Claude Sonnet 5, the Gemini 3 series, and Grok 4.5; Max adds Claude Opus 4.8 and Claude Fable 5 as well.

The list of models is updated frequently, so the catalog should not treat any particular model as a permanent asset. Different models vary in terms of speed, context handling, reasoning capabilities, and resource consumption; the actual availability of a model is determined by the information on its account page.

OCR and image recognition

Pro offers OCR and semantic image recognition for scanning text in papers, screenshots, and charts. OCR results may misinterpret subscripts, mathematical symbols, and multi-column formatting; therefore, key data must be checked against the original image.

Price comparison of Unriddle and Anara packages

The official website displays prices that have been localized according to the country and currency, and it offers monthly and annual payment options as well as regional discounts; annual payment can save around 20% at present.

The exact amount, taxes, as well as information regarding trials and renewals shall be based on the settlement page after logging in.

Unriddle (now Anara) usage guide

Complete a basic task.

1. Clarify the issue, time frame, location, source priority, and output format;
2. Upload materials for which you have permission to use, or enter search queries in Unriddle (now Anara);
3. First, establish a framework by uploading papers and PDFs for reading, and then supplement the evidence through multi-document question answering.
4. It is necessary to distinguish between factual information from the source, the author’s opinions, and AI-generated conclusions.
5. Check each item for dates, numbers, the original location, and any conflicting evidence;
6. The conclusions are manually revised, the verification time is recorded, and then they are published;

Create reusable professional workflows

1. Break down complex topics into four categories of questions: background, data, comparison, and conclusions;
2. Uploading papers in combination with PDF reading, as well as multi-document Q&A and natural language search in databases, constitute the standard steps in research.
3. Give priority to using the official website, research papers, regulatory documents, and raw data;
4. A second person is assigned to review conclusions that are considered high-risk;
5. Save queries, evidence, versions, and unresolved issues;
6. Re-run after the data changes and update the conclusions;

Limits of the free version

• According to the official FAQs, Free version does not require a credit card; users can upload documents and receive responses with citations, with a limit of 2000 AI words per document and 5 uploads per day.
• The free quota is suitable for trying things out, but it’s easy to reach the limit when reading a large number of papers in succession.

Refunds and cancellations

Users can cancel their subscription on the billing settings page, and no fee is charged for such cancellation. The paid benefits generally remain valid until the end of the current period. The provider currently offers a 7-day grace period after the first payment.

The specific eligibility for a refund should be verified through in-app support.

Language support

The Anara interface supports more than 30 languages, including Chinese; conversations can be conducted in over 90 languages, with responses typically given in the language used by the user.

Multilingual skills do not equate to professional translation certification; terms and quotations still need to be checked manually.

Desktop and web versions

The product offers a Web workspace, and it also provides a link for downloading the Anara desktop version. The desktop application remains linked to cloud services and account benefits; it does not enable fully offline operation or local privatization.

Privacy and security

According to the official statements, Anara and its partner AI providers do not use user data to train models, nor do they sell or share such data with third parties; users can delete their files. Enterprise also offers solutions that comply with SSO, SAML, SCIM, SOC 2 Type II, ISO 27001, and HIPAA standards.

Before handling unpublished papers, patient data, and trade secrets, the organization should still verify the DPA, sub-processors, data retention and deletion procedures, as well as the applicable contracts; compliance decisions cannot be based solely on the product pages.

GitHub and the open-source status

Anara has made Lector available on its official GitHub repository; this is a fundamental component used for developing PDF viewers, and it is maintained by Anara Labs. The fact that Lector is open source does not mean that Anara’s AI research platform, search backend, model orchestration tools, and account management services are also open source.

Therefore, the catalog should indicate that \"the core platform is closed-source, while some PDF viewing components are open-source\"; it cannot be claimed that Unriddle or Anara can be deployed in a fully private manner.

Which users is it suitable for?

• Students and researchers who need to read and organize papers quickly;
• A team responsible for conducting literature reviews, comparing different papers, and synthesizing evidence;
• Users who are already using Zotero or Mendeley to manage their references;
• Researchers who need access to biomedical databases;
• Professional teams that wish to enable unified searching of cloud storage, notes, and research materials;
• Organizations that place emphasis on traceable citations and institutional security controls.

Product advantages

• Answers and writings emphasize the citation of sources and verification of the original text;
• It supports comprehensive Q&A for single papers as well as multiple papers;
• Search personal databases, academic databases, and web pages simultaneously;
• Built-in notes, annotations, citations, and knowledge graphs;
• Connect Zotero, Mendeley, cloud storage services, and research databases;
• It offers a free version as well as individual, advanced research, and enterprise packages.

Restrictions and Precautions

• The product has been renamed, and the Unriddle interface as well as the packages mentioned in the old tutorials may be outdated.
• The current price is displayed dynamically based on the region;
• AI may miss documents, misinterpret tables, confuse correlation with causation, or make errors in matching citations.
• A rigorous study cannot do without reading the original text, searching databases, and evaluating empirical evidence;
• When it comes to academic publications, the policies regarding the use and disclosure of AI must also be followed.

Frequently Asked Questions

Has Unriddle changed its name?

Yes. Unriddle’s current brand name is Anara, but the old entry points and historical materials may still show Unriddle.

Is Unriddle free?

There is a Free plan that does not require a credit card; it includes up to 2,000 AI words and 5 uploads per day.

Can multiple PDFs be asked about at the same time?

Yes. By placing documents in the same database, folder, or chat, it is possible to perform cross-file searches and comparisons.

Is Zotero supported?

Supported; the Plus plan and higher offer integration with Zotero and Mendeley.

Is Anara open source?

The core platform is not open source; the official team has made available the Lector PDF viewing component, but it does not include the source code for the full AI-related services.