# Terse

> Terse is a macOS token optimization tool that reduces AI prompt costs by 40-70% through intelligent compression, duplicate tool call detection, redundant file read flagging, and real-time per-turn cost tracking across Claude Code, Cursor, ChatGPT, OpenClaw, and Aider. It runs 100% on-device with zero cloud dependency.

## Key Facts
- Reduces token usage by 40-70% on verbose prompts and up to 89% on CLI output
- 20+ optimization techniques in a 7-stage pipeline
- Auto-detects Claude Code, Cursor, OpenClaw, Aider agent sessions
- Model-aware pricing: tracks costs for Opus, Sonnet, Haiku, GPT-4o, Gemini
- 400+ built-in typo corrections, 130+ phrase-shortening rules
- 3 modes: Soft (typo-fix only), Normal (filler removal), Aggressive (max compression)
- Pricing: Free (50/week), Pro $7.99/mo (unlimited), Premium $99/mo (teams)
- macOS only, built with Tauri + Rust + Swift

## Documentation
- [Getting Started](https://www.terseai.org/docs): Complete usage guide, installation, and configuration
- [What Is Token Optimization?](https://www.terseai.org/what-is-token-optimization): Full guide to token optimization concepts and techniques
- [How to Reduce AI API Costs](https://www.terseai.org/reduce-ai-api-costs): 7 proven strategies to cut AI costs by 60-80%
- [AI Token Pricing Comparison 2026](https://www.terseai.org/ai-token-pricing-comparison): Claude vs GPT vs Gemini complete cost guide

## Techniques
- [Spell Correction](https://www.terseai.org/spell-correction): Hybrid dictionary + macOS NSSpellChecker pipeline
- [Pattern Optimization](https://www.terseai.org/pattern-optimization): 130+ phrase-shortening and redundancy rules
- [NLP Analysis](https://www.terseai.org/nlp-analysis): Filler, hedging, and meta-language removal
- [Telegraph Compression](https://www.terseai.org/telegraph-compression): Aggressive article/preposition stripping

## Research
- [LLMLingua (EMNLP 2023)](https://www.terseai.org/llmlingua): How PruneAI builds on academic prompt compression research
- [Norvig Spelling](https://www.terseai.org/norvig-spelling): Statistical spelling correction foundation
- [Selective Context](https://www.terseai.org/selective-context): Context pruning for token reduction

## Pricing
- Free: 50 optimizations/week, 1 session, all 3 modes, spell correction
- Pro ($7.99/mo): Unlimited optimizations, 3 sessions, agent monitoring, CLAUDE.md generation
- Premium ($99/mo): Unlimited everything, unlimited devices, priority support