Overview

Synapulse (Synapse + Pulse) is a self-hosted personal AI assistant that lives in your Discord server. The idea came from OpenClaw — after seeing what it could do, I decided to build the personal assistant I had always wanted, one that is lightweight, transparent, and fully under my control.

• GitHub: GOODDAYDAY/Synapulse

Demo

Weather query	Web search + recommendation

Reminder (notify mode)	Reminder (prompt mode)

File creation + send

Features

Architecture

Tech Stack

Project Structure

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

apps/bot/
├── main.py                    # Entry point
├── config/                    # Settings, prompts, logging
├── core/
│   ├── handler.py             # Bootstrap: wire all components via DI
│   ├── loader.py              # Auto-discover tools and jobs from folders
│   ├── mention.py             # Tool-call loop, memory load/save, summarization
│   └── reminder.py            # Background reminder checker
config/                          # Runtime config (models.yaml, mcp.json, jobs.json)
output/                          # Runtime output (logs, data)
├── provider/
│   ├── base.py                # OpenAIProvider, AnthropicProvider
│   ├── endpoint.py            # EndpointPool: rotation, rate-limit fallback
│   └── copilot/auth.py        # GitHub OAuth Device Flow
├── channel/
│   └── discord/client.py      # Discord bot integration
├── tool/                      # Native tools (auto-discovered)
│   ├── base.py                # BaseTool, OpenAITool, AnthropicTool
│   ├── brave_search/          # Web search
│   ├── local_files/           # Read/write local files (sandboxed)
│   ├── memo/                  # Notes management
│   ├── task/                  # Todo list
│   ├── reminder/              # Reminders
│   ├── weather/               # Weather via OpenWeatherMap
│   ├── shell_exec/            # Shell commands (with safety blacklist)
│   └── mcp_server/            # Manage MCP connections via chat
├── job/                       # Background jobs (auto-discovered)
│   ├── cron.py                # CronJob base with hot-reload
│   ├── gmail/                 # Gmail monitoring
│   ├── outlook/               # Outlook monitoring
│   └── qqmail/                # QQ Mail monitoring
├── mcp/
│   └── client.py              # MCPManager: spawn, discover, route
└── memory/
    └── database.py            # JSON file persistence

Core Loop

The tool-call loop in core/mention.py:

Overview

ChromePilot is a Chrome extension that lets you control any webpage using natural language. Type a command like “click the login button” or “fill in my email”, and ChromePilot executes it automatically — clicking, typing, scrolling, and navigating on your behalf.

• Built with AI (Claude) assistance: 3 hours for the initial prototype, 5 hours to polish into v1.0
• Current status: v1.0 — functional and usable, with room for further optimization
• GitHub: GOODDAYDAY/ChromePilot

Features

Demo

Basic Actions — Click Repetition

Command: “drink water 10 times”

Overview

• MySQL is the most widely used relational database in the world.
• As business grows, a single MySQL instance becomes the bottleneck — both in capacity and reliability.
• This blog traces the evolution of MySQL high availability: from simple replication to consensus-based clustering.

Evolution History: The Long Road to High Availability

Single Node Era

The starting point. One MySQL instance handles everything.

Overview

• The common three cluster modes are Redis(Split Cluster), MySQL(Master-Slave), Kafka(RAFT).
• This blog will introduce from the single node to the cluster mode of Redis.

Evolution History: From Single Node to Cluster

Single Node Era

• This is the common monolithic architecture of Redis.

Master-Slave Replication

• As to reliability, it is easy to think adding some nodes as Master-Slave.

Introduction

What is Dynamic Programming?

Dynamic Programming (DP) is often considered one of the most challenging topics in computer science algorithms. However, at its core, it is simply an optimization technique.

The fundamental idea of DP is “Don’t Repeat Yourself.”

If you have already solved a sub-problem, you should save the result (cache it) so that you never have to calculate it again. By trading a little bit of space (to store results) for time (to avoid re-calculation), DP can turn an inefficient exponential algorithm ($O(2^n)$) into a highly efficient linear one ($O(n)$).

Preface

When handling high-concurrency scenarios in Spring applications, proper use of asynchronous programming and thread pool management is crucial. This article provides an in-depth analysis of Spring’s default thread pool, custom thread pools, and thread reuse mechanisms through practical code examples.

Why Use Custom Thread Pools?

When a Spring Boot application starts, it automatically configures a global task executor (TaskExecutor) with the default name applicationTaskExecutor. However, using Spring’s default thread pool directly in production environments is not recommended for the following reasons: