save-tokens

Save 94% on AI coding tokens. Index your codebase, agents search instead of reading files. Works with Claude Code, Codex, Copilot, Cursor, Gemini CLI. Local MCP server, free, open source.

Architectural enhancement suite for Hermes Agent: vector memory, adaptive soul, and workflow classification.

Deterministic context compressor for MCP agents. Slots in front of prompt caching, mem0, Letta, or Claude Code. Single Rust binary, zero ML at query time.

一个可复用的 DSH Agent Preset：双阶段、Token 精益的编码智能体模式。 核心目标：把模型思维链的起手习惯从 "let me" 翻转为 "we need"，同时把上下文占用与 Token 消耗压到最低，并保持供应商 KV Cache 对未变前缀持续命中。

Runaway-agent kill switch for AI coding sessions. Detects retry loops, pauses token burn, and alerts operators in real time.

Resume any coding session in a new AI agent for <500 tokens, not 20k+ - automatic capture via git hooks, zero LLM calls, works with any MCP-capable agent.

A macOS clipboard for AI agents that reduces token usage by 29 to 59% on every image you copy.

Fork instead of reinvent. GitHub search skill that finds the best starting points for your new projects, so you start from 80%, not zero.