LLM Cost Optimizer for OpenClaw

English | 中文

---

English

Status: Alpha (core design complete, implementation in progress) | License: MIT | Author: maichanks

⚡ One-Click Deploy: curl -fsSL https://raw.githubusercontent.com/maichanks/llm-cost-optimizer/main/deploy.js -o deploy.js && node deploy.js

Monitor, analyze, and reduce LLM API costs across providers. Features planned: token tracking, prompt compression, semantic caching, budget alerts. Currently a design placeholder — deploy to see development progress and roadmap.

📊 Architecture ready | 🆓 MIT | 🔜 In development | ⭐ OpenClaw compatible

Features

• Token usage tracking per model
• Prompt compression heuristics
• Redis-based semantic caching
• Budget alerts (Slack/Feishu)
• CSV reports
• OpenClaw integration: send alerts to OpenClaw channels

🚀 One-Click Deploy

Run the automated deployment script:

curl -fsSL https://raw.githubusercontent.com/maichanks/llm-cost-optimizer/main/deploy.js -o deploy.js && node deploy.js

This will clone the repo, install dependencies, and create a .env file.

---

Quick Start (Standalone)

git clone https://github.com/maichanks/llm-cost-optimizer.git
cd llm-cost-optimizer
pnpm install
cp .env.example .env
# Edit .env with your keys and alert settings
node scripts/analyze.js --input ./logs
node monitor.js

OpenClaw Integration

You can run this as a background service and send budget alerts through OpenClaw:

1. Install to OpenClaw skills (optional, for unified management):

cp -r llm-cost-optimizer $HOME/.openclaw/workspace/skills/
cd $HOME/.openclaw/workspace/skills/llm-cost-optimizer
pnpm install

2. Configure alert channel in .env:

ALERT_CHANNEL=feishu    # or telegram, slack
ALERT_TARGET=ou_OPEN_ID # your OpenClaw user/chat open_id

3. Add cron for continuous monitoring (runs in isolated session):

openclaw cron add \
  --name "LLM Cost Monitor" \
  --cron "*/30 * * * *" \
  --session isolated \
  --message "node $HOME/.openclaw/workspace/skills/llm-cost-optimizer/monitor.js"

The monitor.js script will check usage and send alerts via OpenClaw message API when budget thresholds are exceeded.

---

中文

状态: Alpha | 许可证: MIT | 作者: maichanks

用于监控、分析和降低多模型 LLM API 成本的实用工具集，可独立运行或与 OpenClaw 集成。

功能

• 按模型统计 token 消耗
• 提示词压缩启发式规则
• 基于 Redis 的语义缓存
• 预算告警（支持 Slack/飞书）
• CSV 报表导出
• OpenClaw 集成：告警可通过 OpenClaw 频道发送

快速开始（独立模式）

git clone https://github.com/maichanks/llm-cost-optimizer.git
cd llm-cost-optimizer
pnpm install
cp .env.example .env
# 编辑 .env 填入密钥和告警设置
node scripts/analyze.js --input ./logs
node monitor.js

OpenClaw 集成

可将此工具作为后台服务运行，并通过 OpenClaw 发送预算告警：

1. 安装到 OpenClaw skills（可选，统一管理）：

cp -r llm-cost-optimizer $HOME/.openclaw/workspace/skills/
cd $HOME/.openclaw/workspace/skills/llm-cost-optimizer
pnpm install

2. 配置告警频道 在 .env 中：

ALERT_CHANNEL=feishu    # 或 telegram, slack
ALERT_TARGET=ou_OPEN_ID # 你的 OpenClaw 用户/群组 open_id

3. 添加 cron 定时监控（在独立会话运行）：

openclaw cron add \
  --name "LLM Cost Monitor" \
  --cron "*/30 * * * *" \
  --session isolated \
  --message "node $HOME/.openclaw/workspace/skills/llm-cost-optimizer/monitor.js"

monitor.js 脚本会检查使用情况，当超出预算阈值时通过 OpenClaw 消息 API 发送告警。

---

📝 Keywords

openclaw, llm, cost-optimization, openrouter, budget-monitoring, token-tracking, semantic-caching, redis, prompt-compression, ai-cost, api-cost-management

---

🔗 Related OpenClaw Projects

• Smart Digest - AI-powered news digest for OpenClaw
• OpenClaw GitHub Trending Notifier - GitHub trending notifications
• Security Hardening for OpenClaw - Security toolkit
• Multi-Platform Publisher - Multi-platform content publishing

---

📄 License

MIT © 2026 maichanks hankan1993@gmail.com