CorpusCraft Documentation

Creating an Account

CorpusCraft uses magic link authentication for secure, passwordless access:

1. Click "Get Started" on the homepage
2. Enter your email address
3. Check your email for the login link
4. Click the link to access your account

Creating Your First Corpus

1. From your dashboard, click "Create New Corpus"
2. Enter a name and optional description
3. Click "Create Corpus"
4. Start uploading documents

Pricing Tiers

CorpusCraft offers five pricing tiers designed for different research needs:

Verification Status Badge

Once verified, you'll see a green verification badge on your dashboard and profile page showing:

• Verified Status: Green badge with checkmark
• Verification Method: Email Domain, ORCID, or Manual Review
• Pending Review: Yellow badge while awaiting manual review

Supported File Formats

• TXT - Plain text files
• CSV - Comma-separated values (with text column)
• JSONL - JSON Lines format for structured data
• PDF - Portable Document Format
• DOCX - Microsoft Word documents

Uploading Documents

Navigate to the "Upload" tab in your corpus and either:

• Click to browse and select files
• Drag and drop files into the upload area

Documents are automatically analyzed for language detection.

Create from Text

Create documents directly by writing or pasting text:

1. Navigate to your corpus page
2. Find the "Create from Text" section
3. Enter a document title
4. Write or paste your text content
5. Click "Create Document" to add it to your corpus

This is perfect for adding quick notes, manually transcribed content, or pasting text from other sources.

Full-Text Search

CorpusCraft uses SQLite FTS5 for powerful full-text search with operators:

• guerra - Find exact word
• "guerra civil" - Phrase search
• guerra OR paz - Either word
• guerra NOT civil - Exclude term
• guerr* - Prefix matching (guerra, guerras, etc.)

KWIC Concordance

Key Word in Context (KWIC) displays search results with surrounding context:

• Left Context - Words before the keyword
• Keyword - Your search term (highlighted)
• Right Context - Words after the keyword
• Source - Document title

Click column headers to sort by left context, keyword, right context, or source.

Context Window

Adjust the number of words shown before and after the keyword (5-20 words).

Regex Search Advanced

Enable regex mode in Advanced Search for pattern-based searching:

• \b(war|battle|conflict)\b - Word boundaries for alternative words
• \d{4} - Match 4-digit numbers (e.g., years)
• colou?r - Optional characters (color or colour)
• [A-Z][a-z]+ - Capitalized words
• \b\w{10,}\b - Words with 10+ characters

Batch Operations Bulk Editing

Perform operations on multiple documents at once:

• Bulk Delete - Remove multiple documents in one action
• Metadata Update - Apply same metadata value to selected documents
• Bulk Export - Export selected documents as TXT or CSV

Word Frequency

Analyze how often words appear in your corpus:

• Set minimum frequency threshold to filter rare words
• View frequency counts and percentages
• Results are sortable by rank, word, or frequency
• Export results to PDF, Word, Excel, or CSV

N-gram Analysis

Analyze multi-word sequences:

• Bigrams - 2-word sequences (e.g., "guerra civil")
• Trigrams - 3-word sequences
• 4-grams+ - Longer sequences

Collocation Analysis

Find words that frequently appear together with a target word using multiple statistical measures:

• MI score - Mutual Information, measures association strength
• t-score - Frequency-sensitive, good for common words
• Dice coefficient - Symmetrical association (0-1 scale)
• Log-Likelihood - Statistical significance measure

Readability Indices Free

Assess text complexity using multiple readability formulas:

• Flesch Reading Ease - 0-100 scale, higher = easier
• Flesch-Kincaid Grade Level - US grade level required
• Gunning Fog Index - Years of education needed
• SMOG Index - Simple Measure of Gobbledygook
• Coleman-Liau Index - Character-based grade level
• Automated Readability Index (ARI) - Character-based formula

Lexical Diversity Researcher+

Measure vocabulary richness using advanced metrics:

• Type-Token Ratio (TTR) - Simple vocabulary diversity measure
• Standardized TTR (STTR) - Less affected by text length
• MTLD - Measure of Textual Lexical Diversity, most stable metric
• HD-D - Hypergeometric Distribution D, probabilistic approach
• Yule's K - Vocabulary concentration measure
• Hapax Legomena - Words appearing only once

Keyness Analysis Professional+

Compare two corpora to identify statistically significant differences:

• Chi-square test - Statistical significance testing
• Log-likelihood - More reliable for sparse data
• Effect size - Practical significance (log ratio)
• Normalized frequencies - Per 10,000 words comparison

Document Analysis

Search & Discovery

Stylistic Analysis

Linguistic Analysis

Comparative Analysis

How to Use NLP Processing

1. Navigate to your corpus page
2. Click the NLP tab in the analysis navigation
3. Select a document from the dropdown menu
4. Click "Analyze Document" to process with spaCy

Supported Languages

NLP processing is available for 8 languages:

• English - en_core_web_sm (advanced pipeline with lemmatizer, NER, parser)
• Spanish - es_core_news_sm (optimized for Spanish text)
• Russian - ru_core_news_sm (Cyrillic text support)
• French - fr_core_news_sm (complete French pipeline)
• German - de_core_news_sm (German morphology support)
• Chinese - zh_core_web_sm (simplified Chinese with word segmentation)
• Japanese - ja_core_news_sm (Japanese tokenization and analysis)
• Arabic - pyarabic + NLTK (tokenization, stemming, stopwords)

NLP Features

• Complete Token Analysis - Lemmatization, POS tagging, and dependency parsing for all tokens with stopword identification
• Sentence Segmentation - Automatic sentence extraction with statistics (total count, average/min/max length in tokens)
• Noun Chunks - Extract noun phrases with root words and dependency relations
• Morphological Features - Detailed grammatical analysis including gender, case, number, tense, person, mood, and aspect
• Lemma Frequencies - Ranked frequency distribution of base forms excluding stopwords
• Stopword Filtering - Identify and filter function words to focus on content words
• Named Entity Recognition - Identify people (PERSON), places (GPE, LOC), organizations (ORG), dates, and more
• Dependency Parsing - Analyze grammatical structure and relationships between words

Analysis Results

The NLP tab displays comprehensive results including:

• Document Statistics - Total tokens, tokens without stopwords, language information
• Sentence Statistics - Total sentences, average/min/max sentence length with full sentence list
• Noun Chunks - All extracted noun phrases with root words and dependencies
• Lemma Frequencies - Complete ranked frequency list excluding stopwords with counts and percentages
• Morphological Features - All tokens with grammatical features (gender, case, tense, etc.)
• POS Distribution - All part-of-speech tags with counts
• Named Entities - All detected entities with their labels
• Complete Token Analysis - All tokens showing text, lemma, POS tag, fine-grained tag, dependency relation, and stopword status
• Professional Exports - Export complete analysis to PDF, Word, Excel, or CSV formats

Available Visualizations

• Word Clouds - Visual representation of word frequencies
• Frequency Distribution - Bar charts showing word frequencies
• Lexical Dispersion - Track word usage across documents

Customization

Most visualizations allow you to adjust parameters like word count, colors, and filtering.

Export Formats

All analysis results can be exported in multiple formats:

• PDF - Professional reports with formatted tables
• Word (DOCX) - Editable documents for further work
• Excel (XLSX) - Spreadsheets for data analysis
• CSV - Import into other tools

Auto-Save Feature

All analysis results are automatically saved for later export:

• KWIC concordance results
• Frequency analysis
• All 18 AI analysis features
• NLP processing results

Exports Page

Access all your saved outputs from the Exports page:

• Browse all saved analysis results
• Filter by corpus or analysis type
• Export individual results in any format
• Bulk export: select multiple outputs and combine into one document

Corpus Export

Export your entire corpus as JSONL for backup or sharing.

Adding Collaborators

Share your corpus with other researchers:

1. Open the corpus you want to share
2. Go to the "Collaborators" section
3. Add collaborators by email
4. Set permissions (view, edit, admin)

Annotation Layers

Create annotation layers for collaborative markup and analysis.

Creating Snapshots

Save versions of your corpus at different stages:

1. Go to the "Export & Snapshots" section
2. Click "Create Snapshot"
3. Add a tag (e.g., "v1.0", "pre-cleanup")
4. Review your snapshot history

Restoring Snapshots

Snapshots can be used to restore your corpus to a previous state if needed.