SparkEngineAI

SparkEngineAI / QuantClaw-plugin

Public

QuantClaw is a plug-and-play task-type routing quantization plugin for OpenClaw.

sparkengineai.github.ioQuantClaw agents claude codex harness llm
39
0
100% credibility
Found Apr 22, 2026 at 39 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

QuantClaw is an OpenClaw plugin that automatically classifies user requests by task type and routes them to appropriate low-precision or high-precision AI models to balance quality, speed, and cost.

How It Works

1
🔍 Discover QuantClaw

You hear about QuantClaw, a smart helper that makes your AI assistant pick the right speed and smarts for different questions to save time and money.

2
🛠️ Add it to your AI setup

You easily add QuantClaw to your existing AI chat system with a simple command, and it creates a starter guide for you.

3
🧠 Tell it your task preferences

You describe simple rules like 'use top quality for coding, fast mode for quick facts' so it knows when to go slow and careful or quick and cheap.

4
📊 Open the dashboard

You visit a friendly web page to see live stats, costs, and tweak settings on the fly as it learns from your chats.

5
💬 Chat naturally

You start asking questions, and QuantClaw quietly sorts each one to the best fit, feeling seamless and smart.

🎉 Smarter, faster AI

Your assistant now delivers great answers quicker and cheaper, with a dashboard showing exactly how much you save.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 39 to 39 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is QuantClaw-plugin?

QuantClaw is a plug-and-play TypeScript plugin for OpenClaw that automates task-type routing quantization. It classifies incoming requests—like coding, safety checks, or data analysis—and routes them to 4bit, 8bit, or 16bit model targets, optimizing quality, latency, and cost without manual user tweaks. Install via `openclaw plugins install @sparkengineai/quantclaw`, tweak `~/.openclaw/quantclaw.json`, and track everything on the built-in dashboard at `/plugins/quantclaw/stats`.

Why is it gaining traction?

It stands out with quantization tuned on real OpenClaw workloads via Claw-Eval benchmarks, proving low-precision works for QA or lookups while reserving high-precision for sensitive tasks. Developers dig the hot-reload config, rule-based or embedding/LLM detectors, and live telemetry for tokens, costs, and sessions—no black box. The dashboard lets you test classifications dry-run and override per-session.

Who should use this?

OpenClaw server runners handling mixed inference loads, like AI agents blending research queries with code gen or safety evals. Prod teams at startups scaling multi-model fleets on quantized backends (e.g., GLM or Qwen variants) who need cost telemetry without custom routers.

Verdict

Grab it if you're on OpenClaw and want hands-off quantization routing—docs shine, MIT license, dashboard accelerates iteration. Low 39 stars and 1.0% credibility score flag early maturity; validate on your workloads before prime time.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.