arXiv & BibTeX

Overview

TeX64 integrates arXiv paper search and BibTeX management directly into the editor, powered by AI. Search for papers, import citations, and manage your bibliography without leaving the editor. These features streamline academic workflows and reduce time spent on literature research and citation management.

Use the AI chat within the editor to search for papers on any topic. The AI returns paper titles, authors, abstracts, and arXiv IDs. Insert citations directly into your .bib file from search results, with metadata automatically formatted in BibTeX form.

arXiv Search

TeX64's AI assistant integrates arXiv search capabilities, allowing you to find papers using natural language queries. Simply describe your request in the AI chat and get back relevant recommendations. The system supports multiple search keywords, date ranges, author names, and category filtering for highly flexible queries.

Basic Search

Request searches in natural language such as:

  • "Find papers about transformer attention mechanisms from 2023"
  • "Latest papers on explainability in machine learning"
  • "arXiv papers on Neural Architecture Search"
  • "Papers about Stable Diffusion or LDM from 2023-2024"
  • "Papers on fine-tuning and efficiency for LLMs"

Search Results

The AI returns comprehensive paper metadata including:

  • Paper title — Exact and searchable format
  • Author list — All authors in BibTeX format
  • Abstract — Summary or first paragraph of the paper
  • arXiv ID — e.g., 2301.12345, retained for future reference
  • Publication date — Submission and latest version dates
  • Categories — Primary and secondary categories (cs.AI, cs.LG, etc.)
  • DOI — Link to published version if available

Importing Citations

One-click workflow to import citations directly from arXiv search results into your .bib file. The AI automatically generates proper BibTeX format and adds the entry to your bibliography.

Direct Import from Search

After presenting search results, the AI supports:

  • Individual entry selection and import
  • Batch import of multiple results
  • Accurate author name and title formatting
  • Automatic preservation of arXiv ID
  • Duplicate checking (warns if already in .bib)

arXiv ID Direct Import

Import specific papers by providing their arXiv ID directly. This is most convenient when you already know the paper ID.

Example:
"Import arXiv ID 2301.12345"

The AI fetches metadata and automatically generates the correct BibTeX entry for your .bib file.

BibTeX Management

TeX64's AI can read and understand your .bib files completely. Ask it to add new entries, fix formatting, check for duplicates, or organize your bibliography. The AI understands complex BibTeX syntax and conventions for efficient citation management.

Supported Entry Types

The AI understands standard BibTeX entry types:

  • @article — Journal papers. For peer-reviewed articles in academic journals.
  • @inproceedings — Conference papers. For papers in conference proceedings.
  • @book — Books. For complete published works by authors.
  • @inbook — Book chapters. For chapters or sections within a book.
  • @incollection — Collection articles. For articles in edited collections.
  • @misc — Miscellaneous. For websites, technical reports, preprints, etc.
  • @thesis — Theses. For master's or doctoral dissertations.
  • @phdthesis — PhD theses. Specifically for doctoral dissertations.
  • @mastersthesis — Master's theses. Specifically for master's dissertations.

BibTeX Fields

The AI fully understands BibTeX field syntax and conventions, supporting these standard fields:

  • title — Title of the publication (often required)
  • author — Author list (in "Author1 and Author2" format)
  • year — Year of publication (or first submission year)
  • journal — Journal name (for @article)
  • booktitle — Conference or book title (for @inproceedings)
  • publisher — Publisher name (for @book)
  • volume — Volume number (of journal or book series)
  • number — Issue number (for journals)
  • pages — Page range (e.g., "123--145")
  • doi — Digital Object Identifier
  • url — URL of online resource
  • arxivId — arXiv ID (TeX64 extension field)
  • abstract — Paper abstract (for reference)

Citation Key Management

Citation keys are unique identifiers used to reference bibliography entries in your TeX document. The AI supports citation key naming conventions, can check consistency, auto-generate, and fix them.

  • Recommended format: author-year (e.g., smith2023, johnson-et-al-2022)
  • AI auto-generates keys from paper authors and year
  • Consistency checking: detects duplicate keys for same authors
  • Auto-matching with \cite references in your TeX source
  • Preservation options: AI suggests changes but you confirm
Tip: Citation keys you use in your paper help the AI automatically detect which references you're missing.

Duplicate Detection and Merging

The AI automatically detects duplicate entries in your .bib file. When the same paper is listed multiple times with different citation keys, the AI proposes merging them.

  • Exact match detection via title and authors
  • Partial match identification (e.g., typos)
  • Auto-matching via arXiv ID
  • Duplicate detection via DOI
  • Retains most complete metadata when merging
  • You can choose which citation key to keep

AI Citation Workflow

Describe what you need in natural language and the AI handles the search and formatting. Focus on your academic work without worrying about implementation details.

Practical Examples

Example 1: Topic-based Search

User:

"Find papers about Vision Transformers from 2022-2024 and add them to my .bib file"

AI does:

  • Searches arXiv for "Vision Transformers"
  • Filters for relevant papers from 2022–2024
  • Presents top results (typically 5–10 papers)
  • Generates BibTeX entries for each paper
  • Adds them to your .bib file after your approval
Example 2: arXiv ID Import

User:

"Add BibTeX entry for arXiv:2301.12345"

AI does:

  • Retrieves metadata from arXiv
  • Parses title, authors, abstract, and date
  • Generates optimal citation key
  • Shows you the formatted BibTeX entry
  • Adds it to your .bib file after approval
Example 3: Auto-fixing Citation Errors

User:

"Here's my compile error. Please fix the undefined citation key"

AI does:

  • Reads compile log
  • Extracts smith2023 from error
  • Checks context around citation in your TeX source
  • Searches arXiv for relevant papers
  • Suggests best matching BibTeX entry
  • Adds it to your .bib file after confirmation

.bib File Operations

TeX64's AI can read your existing .bib file as context. It understands citation keys, entry types, and BibTeX fields, suggesting and applying changes to keep your .bib file clean and organized.

.bib File Analysis

The AI can perform tasks such as:

  • Citation key consistency checks
  • Identifying and merging duplicate entries
  • Detecting formatting inconsistencies
  • Identifying outdated entries for updating
  • Categorizing and organizing entries
  • Identifying missing important fields
  • Validating URL references
  • Standardizing DOI references

Diff and Apply Workflow

You can review proposed changes before applying them. The AI shows you a diff and you approve before updating your .bib file. This protects you from unintended modifications.

Best Practice: Always review large changes in diff format before applying. The AI is accurate, but verify important modifications before approval.

Compile Log Integration

When your build fails due to undefined citations, the AI reads the compile log, identifies the missing citation keys, and can search arXiv to find and add the correct entries.

Auto-repair Workflow

  • TeX compiler outputs error (e.g., "Citation key smith2023 is undefined")
  • Paste compile log to AI chat, ask to fix the error
  • AI parses log and extracts all undefined citation keys
  • Checks your TeX source for context around missing citations
  • Searches arXiv or other sources for relevant papers
  • Proposes BibTeX entries for each undefined key
  • Adds entries to your .bib file after your approval
  • Recompile to verify fixes (new entries should resolve errors)

Undefined Citation Detection

The AI detects \cite references in your TeX source and checks if corresponding BibTeX entries exist. It automatically identifies missing keys and can add them to prevent compile errors.

Tip: The AI can infer which references you need by reading your full TeX document. The more context you provide, the better its recommendations.

natbib/biblatex Compatibility

TeX64 supports both natbib and biblatex bibliography management systems. The AI understands citation commands and .bib file formats for both systems.

natbib Citations

  • \cite{key} — Basic citation
  • \citet{key} — Textual citation
  • \citep{key} — Parenthetical citation
  • \citealt{key} — Citation without parentheses

biblatex Citations

  • \cite{key} — Basic citation
  • \parencite{key} — Parenthetical citation
  • \textcite{key} — Textual citation
  • \fullcite{key} — Full citation information

The AI automatically detects which system you use from your TeX source and proposes citation formats accordingly.

Best Practices

  • Clear Citation Key Naming: Use author-year format (e.g., smith2023, johnson2022). This makes keys easy for the AI to recognize and complete. For multiple authors, use "johnson-et-al-2023" format.
  • Regular Syncing: After adding new papers, have the AI check your .bib file for duplicates and consistency. Weekly checks are recommended best practice.
  • Preserve arXiv IDs: Include arXiv IDs in your .bib file so the AI can easily verify and update entries later. Add "arxivId = 2301.12345" field to each entry for future reference.
  • Share Error Logs: When citation errors occur, share your compile log and relevant TeX source with the AI. Detailed information ensures more accurate repairs.
  • Author Name Formatting: Keep author names in "First Name Last Name and First Name Last Name" format. The AI parses multiple authors correctly and formats them for bibliography style.
  • Verify Before Publishing: Always verify AI-generated BibTeX entries before final publication. Especially check author names, publication years, and journal titles.

Limitations and Notes

  • Search Scope: arXiv search works best for recent papers and popular publications. Very niche topics may have limited results. Older papers (pre-1990s) or specialized domains may have incomplete results.
  • Verification Needed: BibTeX entries generated by the AI are reliable, but always verify before final publication. Especially verify that auto-generated metadata is complete and accurate.
  • Citation Key Inference: Citation keys and author names are most important for AI inference. Without clear specification, the AI recommends the most relevant matches. For multiple authors, the AI uses the first author's surname.
  • Field-Specific Knowledge: The AI is optimized for computer science, physics, and mathematics papers. It works in other fields like medicine and biology, but niche domains may have incomplete metadata.
  • No Retroactive Indexing: The AI does not automatically retroactively fix past compilation errors. It handles current errors. If you have previous compile logs, share them with the AI to request fixes.

Summary

TeX64's arXiv & BibTeX integration makes academic writing more efficient. You can search and manage through natural language AI chat, eliminating the need to remember complex BibTeX syntax. Search papers, import citations, and fix errors without leaving the editor.

Get Started: Open the AI chat in the editor and try: "Find papers about Transformers from 2024". You'll get results in seconds and add them to your .bib file with one click.