Skip to content

charliemarketplace/textreduce

Repository files navigation

TextReduce

Enterprise-grade AI summarization. 100,000+ tokens per second.

The Technology

TextReduce is powered by TextReduce-1, a proprietary neural architecture optimized for extractive semantic distillation. Our model leverages:

  • Adaptive Lexical Saliency Mapping™ — Dynamically computes context-aware term importance across your entire document corpus
  • Sentence-Level Semantic Scoring — Each sentence is evaluated through our multi-pass importance aggregation pipeline
  • Order-Preserving Extraction — Maintains narrative coherence through positional indexing

The result? Blazing-fast summarization that runs entirely on-device—no cloud latency, no API rate limits, no waiting.

"It's like GPT but faster and it runs in a browser" — Internal benchmarks

Website

textreduce.com

  • Paste any text (we've tested up to 50MB documents*)
  • Fine-tune compression ratio with our precision slider
  • Generate shareable links instantly

Zero signup. Zero tracking. Zero inference costs.

*Results may vary. Please don't actually paste 50MB.

API

Production-ready API via URL parameters:

https://textreduce.com/?text=Your+text+here&pct=30
Parameter Description Default
text URL-encoded text to summarize (required)
pct Extraction threshold (0-100) 30
raw Set to 1 for headless mode 0

Examples

# Interactive mode (full UI experience)
https://textreduce.com/?text=Hello.+This+is+a+test.&pct=50

# Headless mode (raw inference output)
https://textreduce.com/?text=Hello.+This+is+a+test.&pct=50&raw=1

TextReduce-1 · On-device AI · No GPU required · Powered by JavaScript

About

100k TPS AI summarization includes production x402 implementation

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors