OPEN SOURCE · AI TOOLKIT

Give your AI
a Library.

Persistent memory, image generation and intelligent token modes — all in one lightweight MCP server and skill bundle.

Get Started → GitHub ↗

⌘ MCP Compatible ✦ Open Source ⚡ Lightweight

sm-ailib / terminal

› sm_ailib_recall

query: "my previous project"

✓ Memory retrieved

› sm_ailib_level

mode: 3 · Standard

✓ Token mode activated

› sm_ailib_img

prompt: "futuristic city"

✓ Image generated

› _

WHAT IT CAN DO

One library. Three superpowers.

Give text-only AI models access to useful local capabilities.

🧠

01 / MEMORY

Persistent Memory

Save important information and recall it later. Your AI can access stored context across sessions.

@sm-ailib_save @sm-ailib_recall

🎨

02 / IMAGES

Image Generation

Generate images from text prompts using Pollinations AI and save the results locally.

@sm_ailib_img

⚡

03 / TOKENS

Token Modes

Switch between six token modes, from strict minimal responses to maximum audit.

@sm_ailib_0 → @sm_ailib_5

CONTROL YOUR CONTEXT

Six modes. Your choice.

Choose how much context your AI should use.

MODE 0

Strict Lazy

256 tokens

Ultra-minimal responses.

@sm_ailib_0

MODE 1

Minimal

2,048 tokens

Minimal senior developer style.

@sm_ailib_1

MODE 2

Focused

4,096 tokens

More context for focused tasks.

@sm_ailib_2

MODE 3 · DEFAULT

Standard

8,192 tokens

Balanced everyday reasoning.

@sm_ailib_3

MODE 4

Maximum Reasoning

16,384 tokens

Expanded context for complex tasks.

@sm_ailib_4

MODE 5

Maximum Audit

32,768 tokens

Maximum context for detailed audits.

@sm_ailib_5

QUICK REFERENCE

Simple commands.

Use these commands to access SM-AILIB capabilities.

@sm-ailib_save <content>

Save information to persistent memory.

@sm-ailib_recall <query>

Search and recall saved information.

@sm_ailib_img <prompt>

Generate an image from a text prompt.

@sm_ailib_0 ... @sm_ailib_5

Switch between the six token modes.

These are skill command formats. The MCP server exposes the corresponding tools to supported AI clients.

GET STARTED

Install SM-AILIB.

Choose your AI coding assistant and follow its installation guide.

Claude Code

Get the repository from GitHub and configure the MCP server.

git clone https://github.com/sgdev-official/SM-AILIB.git

cd SM-AILIB

npm install

Follow the repository's current Claude Code installation instructions to register the MCP server and skill.

Claude Code MCP documentation ↗

OpenCode

Get SM-AILIB from GitHub and configure its OpenCode integration.

git clone https://github.com/sgdev-official/SM-AILIB.git

cd SM-AILIB

npm install

Configure the MCP server, plugin and skill using the OpenCode installation guide in the repository.

OpenCode documentation ↗

View Repository on GitHub ↗

QUESTIONS?

Frequently asked questions.

What is SM-AILIB?

SM-AILIB is an MCP server and local skill bundle that adds persistent memory, image generation and token modes to supported AI clients.

Which AI clients does it support?

It provides integration for OpenCode and Claude-compatible environments. Check the repository for the latest setup details.

Where is my memory stored?

Memory is stored locally by the SM-AILIB server. Check the project documentation for its storage location.

Is SM-AILIB free and open source?

Yes. You can explore the source code and project updates on the public GitHub repository.

Does image generation require an internet connection?

Yes. The image generation feature uses Pollinations AI, so it depends on network availability and that service.

READY TO UPGRADE YOUR AI?

Give your AI a memory.

Start building with SM-AILIB today.

Explore on GitHub ↗

Give your AI a Library.