# Terse

> Terse is an on-device AI agent butler for macOS and Windows that cuts AI token costs by 40-70%. It compresses prompts before they hit the API, live-monitors 8 coding agents (Claude Code, Cursor, OpenAI Codex, GitHub Copilot CLI, Cline, Windsurf, OpenClaw, Aider), enforces spending limits with a budget circuit breaker that can pause or kill a runaway agent before its next API call, manages MCP servers with security risk scoring, and diagnoses token waste with Terse Doctor. It runs 100% on-device — prompts never leave your machine. Also available as a Chrome extension and a VS Code extension.

## Key Facts
- Reduces token usage 40-70% on verbose prompts; 35+ on-device optimization techniques (<5ms)
- Live monitoring of 8 agents: input/output tokens, cache read/write efficiency, per-turn cost (model-aware pricing), burn rate, context-window fill
- Budget circuit breaker: burn-rate / token / dollar ceilings that escalate from alert to pausing (SIGSTOP) or killing (SIGTERM) the agent process before the next API call
- MCP Manager: discovers every MCP server across Claude Code / Cursor / Windsurf configs; risk-scores each (remote transport, embedded credentials, code-execution surface, unpinned supply chain); enable/disable without editing JSON
- Terse Doctor: ~25 consent-gated scans — cache thrash, duplicate tool calls, redundant file reads, context burn, CLAUDE.md token tax, secret exposure, runaway runtimes — with one-click fixes
- 3 optimization modes: Soft (typo-fix only), Normal (filler removal), Aggressive (max compression); code blocks always protected
- Alert Center: macOS/Windows notifications, Slack webhook, ntfy.sh push; weekly savings digest
- Team sharing: publish live agent sessions and aggregate stats to teammates
- Pricing: Free 30-day trial; Monthly $4.99/mo; macOS + Windows, built with Tauri + Rust + Swift

## Pages

### Product
- [Home](https://www.terseai.org/): Full product overview, pricing, features
- [Budget Circuit Breaker](https://www.terseai.org/agent-budget-circuit-breaker): Stop runaway AI agents — spending limits that pause or kill before the next API call
- [MCP Manager](https://www.terseai.org/mcp-manager): Discover, risk-score, and manage MCP servers across your agent configs
- [Terse for Windows](https://www.terseai.org/for-windows): The full Terse app on Windows
- [Token Cost Calculator](https://www.terseai.org/token-calculator): Estimate your potential token savings
- [About Terse](https://www.terseai.org/about): Who builds Terse and why

### Per-Tool Pages
- [Claude Code](https://www.terseai.org/for-claude-code) · [Cursor](https://www.terseai.org/for-cursor) · [ChatGPT](https://www.terseai.org/for-chatgpt) · [GitHub Copilot](https://www.terseai.org/for-github-copilot) · [Aider](https://www.terseai.org/for-aider) · [Cline](https://www.terseai.org/for-cline) · [Windsurf](https://www.terseai.org/for-windsurf) · [Codex CLI](https://www.terseai.org/for-codex-cli) · [Gemini CLI](https://www.terseai.org/for-gemini-cli) · [OpenCode](https://www.terseai.org/for-opencode) · [Zed](https://www.terseai.org/for-zed)

### Comparisons & Alternatives
- [AI coding agent costs (pillar guide)](https://www.terseai.org/ai-coding-agent-costs)
- [Cursor vs Claude Code](https://www.terseai.org/cursor-vs-claude-code) · [Claude Code vs GitHub Copilot](https://www.terseai.org/claude-code-vs-github-copilot) · [Claude Code vs Codex CLI](https://www.terseai.org/claude-code-vs-codex-cli) · [Windsurf vs Claude Code](https://www.terseai.org/windsurf-vs-claude-code) · [Windsurf vs Cursor](https://www.terseai.org/windsurf-vs-cursor) · [Cursor vs GitHub Copilot](https://www.terseai.org/cursor-vs-github-copilot)
- [Claude Code alternatives](https://www.terseai.org/claude-code-alternatives) · [Cursor alternatives](https://www.terseai.org/cursor-alternatives) · [GitHub Copilot alternatives](https://www.terseai.org/github-copilot-alternatives)

### Cost & Pricing Guides
- [Claude Code pricing 2026](https://www.terseai.org/claude-code-pricing) · [Cursor pricing](https://www.terseai.org/cursor-pricing) · [GitHub Copilot pricing](https://www.terseai.org/github-copilot-pricing)
- [AI token pricing comparison](https://www.terseai.org/ai-token-pricing-comparison) · [How much do AI coding agents cost per month](https://www.terseai.org/how-much-do-ai-coding-agents-cost-per-month) · [Cheapest AI coding assistant](https://www.terseai.org/cheapest-ai-coding-assistant)
- [Why is Claude Code so expensive](https://www.terseai.org/why-is-claude-code-so-expensive) · [Why is Cursor so expensive](https://www.terseai.org/why-is-cursor-so-expensive)
- [How to reduce Cursor costs](https://www.terseai.org/how-to-reduce-cursor-costs) · [How to reduce GitHub Copilot costs](https://www.terseai.org/how-to-reduce-github-copilot-costs) · [Reduce AI API costs](https://www.terseai.org/reduce-ai-api-costs)

### Concept Guides
- [What is token optimization](https://www.terseai.org/what-is-token-optimization) · [What is prompt compression](https://www.terseai.org/what-is-prompt-compression) · [What are tokens in LLMs](https://www.terseai.org/what-are-tokens-in-llms) · [What is a context window](https://www.terseai.org/what-is-a-context-window) · [What is context rot](https://www.terseai.org/what-is-context-rot) · [What is an AI coding agent](https://www.terseai.org/what-is-an-ai-coding-agent) · [Input vs output tokens](https://www.terseai.org/input-vs-output-tokens)
- [Token optimization techniques](https://www.terseai.org/token-optimization-techniques) · [How to write prompts that use fewer tokens](https://www.terseai.org/how-to-write-prompts-that-use-fewer-tokens) · [How to use CLAUDE.md to save tokens](https://www.terseai.org/how-to-use-claude-md-to-save-tokens) · [Common mistakes that waste tokens](https://www.terseai.org/common-mistakes-that-waste-tokens)
- [The Terse blog](https://www.terseai.org/blog)

### Technique Deep-Dives
- [LLMLingua](https://www.terseai.org/llmlingua) · [Telegraph compression](https://www.terseai.org/telegraph-compression) · [Pattern optimization](https://www.terseai.org/pattern-optimization) · [NLP analysis](https://www.terseai.org/nlp-analysis) · [Spell correction](https://www.terseai.org/spell-correction) · [Selective context](https://www.terseai.org/selective-context) · [Norvig spelling](https://www.terseai.org/norvig-spelling)

## Pricing
- Free: 30-day trial, all 3 modes, spell correction
- Monthly ($4.99/mo): Unlimited optimizations, agent monitoring, budget circuit breaker, MCP manager, Doctor, CLAUDE.md generation