cntext
Analyzes Chinese and English text using dictionary sentiment, Word2Vec/GloVe embeddings, and WEPA semantic projection.
Measure customer feedback, brand scores, and psychological attitudes from text using custom semantic axes.
cntext is a Python library designed for social science text analysis and semantic measurement. You can use it to clean text, analyze sentiment with custom YAML dictionaries, train Word2Vec or GloVe models, and run Word Embedding Projection Approach (WEPA) scoring. It suits customer experience analysts and researchers who need Python-based text processing on Chinese or English text data.
What it is
The tool processes PDF, DOCX, and raw text files while providing text cleaning features like traditional-to-simplified Chinese character conversion. It constructs semantic axes from theory-driven positive and negative anchor words to project text onto custom psychological or brand dimensions. It also computes readability metrics, sentiment scores, and semantic brand scores.
Who it's for
It suits Python developers and data analysts who want to measure nuanced sentiment, brand perception, or discursive diversity in Chinese customer reviews and support transcripts. It does not suit non-technical teams looking for a visual no-code dashboard or ready-made ticketing interface.
Key features
Word Embedding Projection
Measures construct salience by projecting text onto a semantic axis built from positive and negative anchor words.
Document IO and Cleaning
Reads PDF, DOCX, and text files while offering utilities to convert traditional Chinese characters and fix contractions.
Dictionary Sentiment Scoring
Computes text sentiment using built-in or custom YAML dictionaries across English and Chinese texts.
Embedding Model Utilities
Trains Word2Vec and GloVe embeddings while supporting dictionary expansion and Procrustes alignment.
Up and running in 3 steps
Our honest verdict
What's great
- Provides built-in traditional-to-simplified Chinese conversion and text normalization tools.
- Implements one-line WEPA scoring to project text onto custom positive and negative semantic axes.
- Supports multiple input formats including PDF, DOCX, and raw text files.
- Includes built-in YAML dictionary reading and sentiment scoring functions.
Where it falls short
- Requires Python programming knowledge to write scripts and manage dependencies.
- Requires self-managed computational resources to train custom embedding models.
- Lacks a built-in user interface or visual dashboard for non-technical users.
Pricing & license
cntext is free and open source under the MIT license. The only cost is the computing infrastructure and developer time needed to execute scripts and train custom embeddings.
Open remote AI-training roles
870 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
Use it, or have us set it up
FAQ
Does cntext require an internet connection to run?
No, core functions including text cleaning, dictionary analysis, and WEPA run locally without making network calls.
Can cntext process languages other than Chinese?
Yes, while optimized for Chinese text processing, it supports English language text and custom dictionary formats.
