The on-device AI agent butler. Cut AI coding agent costs 40–70%, live-monitor every agent you run, and stop runaway spend before the next API call.
terseai.org · Main repo → · Compare vs ccusage & others · FAQ
Terse (the tool at terseai.org) watches the AI coding agents you already run — Claude Code, Cursor, OpenAI Codex, GitHub Copilot CLI, Cline, Windsurf, OpenClaw and Aider — and handles the parts that quietly cost money:
| ⚡ | Compress | Every prompt 40–70% smaller before it hits the API. 35+ on-device techniques, sub-5ms, code always protected. |
| 📡 | Monitor | Live tokens, cost, cache read/write efficiency, burn rate and context fill across 8 agents. |
| 🛑 | Stop | A budget circuit breaker that pauses (SIGSTOP) or kills (SIGTERM) a runaway agent before its next API call. |
| 🔌 | Manage | Every MCP server discovered and risk-scored — remote transport, embedded credentials, code-execution surface. |
| 🩺 | Diagnose | ~25 Terse Doctor scans for cache thrash, duplicate tool calls and redundant file reads. |
Everything runs locally. Prompts and sessions never leave your machine.
- terseai — the Terse SDK (MIT), the benchmark harness behind the 40–70% figure, and the product docs.
macOS & Windows · Chrome extension · VS Code extension
terseai.org · Built with Tauri · Rust · Swift