A unified command-line and programmatic tool to manage, start, check, and query local LLM engines and OpenAI-compatible APIs.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install llm-service-manager
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install llm-service-manager using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is LLM Service Manager?
The LLM Service Manager is an essential tool designed to streamline the management of local Large Language Model (LLM) engines and API endpoints. Developers and AI practitioners can easily control local services like Ollama and vLLM, check runtime statuses, start local model hosts, and download models dynamically without leaving their command environment. By integrating this solution with other Openclaw Skills, developers can build responsive, self-healing local AI agent environments that automatically start missing services and load models on demand.
This skill unifies diverse engines—ranging from local Ollama instances on Windows to cloud or self-hosted OpenAI-compatible endpoints—under a single, intuitive execution syntax. It simplifies local AI development by handling backend lifecycle operations programmatically through a clean Python-based runtime interface.
LLM Service Manager Use Cases
- Checking the status of local LLM orchestration runtimes like Ollama and vLLM dynamically.
- Instantly spinning up Ollama services on Windows environments during pipeline execution.
- Automatically pulling required models (such as gemma3:2b) when an agent requires specific capabilities.
- Sending direct queries to local Ollama installations or external OpenAI-compatible APIs through a unified messaging interface.
How LLM Service Manager Works
- Trigger Recognition: The agent identifies key phrases like "LLM", "Ollama", or "本地模型" to invoke the service manager.
- Status Evaluation: Runs a status check across all configured endpoints to see which services are active.
- Lifecycle Orchestration: If a local service is offline, execution commands like "start" or "pull" bring services online or download model weights.
- Query Routing: Translates uniform queries (e.g.,
ollama:<query>orapi:<query>) into targeted backend payloads, returning fast, structured responses.
LLM Service Manager Setup
Prerequisites
Make sure the configuration directories are set up correctly. By default, the skill operates in the local directory:
C:/Users/pc/.config/opencode/skills/llm_service_manager.
Installation
To integrate this with your framework supporting Openclaw Skills, place the skill folder in your designated skills path and install any dependencies.
# Navigate to your skills directory
cd C:/Users/pc/.config/opencode/skills/llm_service_manager
# Run service management commands programmatically in Python
python -c "from llm_service_manager import run; print(run('status'))"
LLM Service Manager Data Schema & Taxonomy
The LLM Service Manager handles commands and routes them to respective engine schemas. Below is a breakdown of the supported command patterns:
| Input Command Pattern | Targeted Backend | Action Description |
|---|---|---|
status |
System / All Services | Evaluates and returns running state of Ollama, vLLM, and API endpoints |
start |
Local OS (Windows) | Boots the background Ollama service daemon |
pull <model> |
Ollama Engine | Triggers model pull and tracks download sequence |
ollama:<query> |
Local Ollama API | Sends direct inference request to local active model |
api:<query> |
OpenAI-Compatible Endpoint | Routes text generation query to external/cloud LLM endpoint |
(any text) |
Auto-detect Routing | Automatically analyzes and dispatches query to the best available engine |
All skill configurations and logs are stored inside the local system path C:/Users/pc/.config/opencode/skills/llm_service_manager.
LLM Service Manager Advanced Features
- Auto-detection Inference: Automatically detects which LLM services are currently active and routes generic text prompts to the optimal online local engine.
- Dynamic Model Fetching: Automates prompt-driven downloading of specific model architectures directly via programmatic command hooks.
- Multi-Engine Support: Seamlessly bridges local execution targets (Ollama/vLLM) with external APIs, creating a resilient hybrid-cloud inference layer inside Openclaw Skills workflows.
SKILL.md
Loading
Related Openclaw Skills
LLM Service Manager
An aggregate knowledge base and automation skill built to optimize Obsidian notes and troubleshoot Windows system services.
Video Generator
An automated programmatic video production pipeline that utilizes React and Remotion to build high-quality motion graphics and cinematic content.
WeChat Rewrite (公众号文案改写)
An AI-powered copywriting agent designed to convert standard text into highly engaging, viral articles tailored specifically for WeChat Official Accounts.
Novel to Audiobook
An AI-powered skill that converts novel text into fully produced multi-voice audiobooks with customized background music.
OpenRouter Model Picker
An interactive, Canvas-based model configuration utility that dynamically fetches all OpenRouter models and applies configuration hot-updates with zero downtime.
Linux Quick Start & Mastery Guide
An AI-curated mastery skill that compiles system administration, RockyLinux setups, embedded engineering, and creative Linux configurations into actionable learning pathways.
Featured*
