什么是 ClawSaver？

ClawSaver 是 Openclaw Skills 生态系统中的高性能工具，旨在消除与冗余 LLM API 调用相关的浪费。在标准智能体架构中，用户的每条后续消息都会触发完整的 API 请求，迫使模型重新处理昂贵的上下文（如系统提示词和聊天历史）。ClawSaver 通过引入智能缓冲层解决了这个问题，该层会等待一瞬间，将相关的输入收集到单个批次中。

通过将多次用户交互合并为一个请求，开发人员可以显著降低 Token 消耗和运营开销。对于希望在保持高质量响应的同时大幅降低每月 API 账单的规模化生产 AI 应用团队来说，这项技能是必不可少的。它被构建为轻量级、安全且易于集成到任何基于 JavaScript 的 AI 工作流中。

下载入口:https://github.com/openclaw/skills/tree/main/skills/ragesaq/clawsaver

安装与下载

1. ClawHub CLI

从源直接安装技能的最快方式。

2. 手动安装

将技能文件夹复制到以下位置之一

优先级：工作区 > 本地 > 内置

3. 提示词安装

将此提示词复制到 OpenClaw 即可自动安装。

Install

ClawSaver 应用场景

• 减少对话类应用中用户发送多条简短后续消息导致的 Token 浪费。
• 优化客户服务机器人，以更具成本效益的方式处理多轮查询。
• 管理 AI 智能体的大量数据摄取，而不会触发过高的单条消息 API 开销。
• 扩展多用户 AI 平台，解决上下文重新加载成本这一主要财务瓶颈。

1. 应用程序接收用户消息并将其路由到 ClawSaver 缓冲区，而不是立即调用 API。
2. 防抖定时器（默认为 800ms）启动，等待来自同一用户会话的任何其他后续消息。
3. 如果在窗口期内有更多消息到达，它们将被添加到队列中并刷新定时器。
4. 一旦缓冲窗口关闭或达到消息上限，系统将刷新并发送整个批次。
5. 向模型发起单次 API 调用，其中包含所有批处理的消息，允许模型仅处理一次上下文。
6. 模型返回综合响应，系统为下一次交互重置。

ClawSaver 配置指南

通过命令行安装该技能：

clawhub install clawsaver

要将 ClawSaver 集成到您的项目中，请在消息处理器中初始化会话防抖器：

import SessionDebouncer from 'clawsaver';

const debouncers = new Map();

function handleMessage(userId, text) {
  if (!debouncers.has(userId)) {
    debouncers.set(userId, new SessionDebouncer(
      userId,
      (msgs) => callModel(userId, msgs)
    ));
  }
  debouncers.get(userId).enqueue({ text });
}

ClawSaver 数据架构与分类体系

ClawSaver 为活动会话维护本地内存状态，以确保零延迟处理。配置和元数据结构如下：

name: clawsaver
description: Reduce model API costs by 20–40% through intelligent message batching. Buffer related messages, send once.
metadata:
  clawdbot:
    emoji: "?"
    requires:
      env: []
      bins: []
    files: ["SessionDebouncer.js", "example-integration.js"]

ClawSaver

Reduce model API costs by 20–40% through intelligent message batching and buffering.

Most agent systems waste money on redundant API calls. When users send follow-up messages, you call the model separately for each one. ClawSaver fixes this by waiting ~800ms to collect related messages, then sending them together in a single optimized request. Same response quality. Lower cost. No user friction.

How It Works: Batching & Buffering

WITHOUT CLAWSAVER (Context Overhead Hidden):
User:  "What is ML?"
Model: → API Call #1 [Context: system prompt, ch@t history] (cost: $X)
       Returns: definition

User:  "Give an example"
Model: → API Call #2 [Context: system prompt, ch@t history, Q1, A1] (cost: $X)
       Returns: example

User:  "Apply to finance?"
Model: → API Call #3 [Context: system prompt, ch@t history, Q1–A2] (cost: $X)
       Returns: finance application

Total: 3 calls × full context = 3X cost, each call repeats context overhead

───────────────────────────────────────

WITH CLAWSAVER (Single Context Load):
User:  "What is ML?"          ← Buffer (800ms wait)
User:  "Give an example"      ← Buffer (800ms wait)
User:  "Apply to finance?"    ← Flush: Send all 3 together

Model: → API Call #1 [Context loaded ONCE: system prompt, ch@t history]
       Processes all 3 questions together
       Returns: comprehensive answer addressing all three

Total: 1 call × full context = 1X cost, context overhead paid once

Actual savings (with context): 67% reduction
Cost per token: 1/3 (fewer context re-loads + consolidation)

Why it matters: Context (system prompts, history, instructions) gets re-sent on every API call. With ClawSaver, you pay that context overhead once per batch instead of three times. This compounds the savings beyond just "fewer calls."

Example (4K token context, 200 output tokens):

• Without ClawSaver: 3 calls × 4,200 tokens = 12,600 tokens
• With ClawSaver: 1 call × 4,600 tokens = 4,600 tokens
• Actual savings: 63% token reduction (even better than call reduction)

The Problem

User: "What is machine learning?"
(pause)
User: "Give an example"
(pause)
User: "How does that apply to healthcare?"

Without optimization: 3 API calls = 3x cost
With ClawSaver: 1 batched call = 1/3 the price

Across thousands of conversations, this compounds fast.

How It Works

1. User sends message → ClawSaver buffers it
2. Waits ~800ms for follow-ups from same user
3. If more messages arrive → keep buffering
4. Timer expires → send all messages together
5. Model responds once → you get complete answer

Why users don't notice: They're already waiting for your model response. Buffering input doesn't feel slower because the response comes right after the batch sends.

Install

clawhub install clawsaver

Quick Start (10 lines)

import SessionDebouncer from 'clawsaver';

const debouncers = new Map();

function handleMessage(userId, text) {
  if (!debouncers.has(userId)) {
    debouncers.set(userId, new SessionDebouncer(
      userId,
      (msgs) => callModel(userId, msgs)
    ));
  }
  debouncers.get(userId).enqueue({ text });
}

Impact

Three Profiles

Choose based on your use case:

Balanced (Default)

• 25–35% savings
• 800ms buffer
• Chat, Q&A, general conversation

Aggressive

• 35–45% savings
• 1.5s buffer
• Batch workflows, high-volume ingestion

Real-Time

• 5–10% savings
• 200ms buffer
• Interactive, voice-first systems

When to Use

? Chat applications
? Customer support bots
? Multi-turn Q&A
? Any conversation with follow-ups

? Single-request workflows
? Sub-100ms response requirements

API

new SessionDebouncer(userId, handler, {
  debounceMs: 800,      // wait time
  maxWaitMs: 3000,      // absolute max
  maxMessages: 5,       // batch size cap
  maxTokens: 2048       // reserved
})

// Methods
debouncer.enqueue(message)      // add to batch
debouncer.forceFlush(reason)    // send now
debouncer.getState()            // buffer + metrics
debouncer.getStatusString()     // human-readable

Docs

• START_HERE.md — Navigation (pick your role/timeline)
• AUTO-INTEGRATION.md — ? Drop-in middleware wrapper (2 min setup)
• QUICKSTART.md — 5-minute integration
• INTEGRATION.md — Patterns, edge cases, full config
• SUMMARY.md — Metrics and ROI (decision makers)
• SKILL.md — Full API reference
• example-integration.js — Copy-paste templates

Security

• No telemetry — Doesn't phone home
• No network calls — Runs locally
• No dependencies — Pure JavaScript
• You control output — You decide what goes to your model

Data never leaves your machine.

License

MIT

Start here: Pick your path in START_HERE.md, or jump to QUICKSTART.md for 5-minute setup.