clawlite

Lightweight agent framework for local LLMs. OpenClaw patterns, none of the weight.

$ npx clawlite
clawlite v0.1.0
Primary: ollama/qwen2.5:32b
Fallback: groq/llama-3.3-70b-versatile
Tools: shell, read_file, write_file, edit_file
Skills: 3 loaded
 
you > find all TODO comments in the codebase
[tool: shell]
 
clawlite > Found 14 TODOs across 8 files...

Why we built this

We run Ziggy, an autonomous AI agent on an NVIDIA DGX Spark. Qwen 2.5 32B, local inference, no cloud. When we looked at OpenClaw for tool-calling and automation, the architecture was sound but the infrastructure was built for cloud-scale deployments.

We needed the patterns without the weight.

Lane-based serial execution. Context window guards. Skill files. Provider fallback. These are the things that make an agent reliable with a small quantised model. The gateway server, channel adapters, web dashboard, and browser automation are not.

ClawLite is what we extracted. ~500 lines of actual logic. Everything OpenClaw got right about agent reliability, nothing it got heavy about.

If you are running a local model and want an agent that can actually use tools without fumbling, this is for you.

What you get

from OpenClaw

Lane-based serial queue

Serial by default, parallel only when you say so. No race conditions. No interleaved garbage.

from OpenClaw

SKILL.md files

Drop markdown in skills/ to shape behaviour. Trigger keywords for dynamic loading. Configure with text, not code.

from OpenClaw

Persistent approvals

Approve a shell command once, it remembers. No re-prompting every session for the same git push.

small LLM optimised

Context window guard

Automatic compaction at 80%. Summarises older turns before you hit the wall. Critical at 16K context.

small LLM optimised

One tool call per turn

Small quants fumble parallel calls. Serial tool execution is more reliable with Qwen, Llama, Mistral.

resilience

Provider fallback

Ollama goes down, Groq API catches it. Automatically. No broken pipelines at 3am.

Get started

git clone https://github.com/ziggybot/clawlite
cd clawlite && npm install
cp config.example.json clawlite.config.json
npm start