InertiaRSS Track and read blogs, news, and tech you care about
Read Original Open in InertiaRSS

Recommended Feeds

V
V2EX
博客园 - 叶小钗
Y
Y Combinator Blog
大猫的无限游戏
大猫的无限游戏
博客园 - 【当耐特】
酷 壳 – CoolShell
酷 壳 – CoolShell
D
Docker
WordPress大学
WordPress大学
Blog — PlanetScale
Blog — PlanetScale
博客园 - Franky
G
Google Developers Blog
爱范儿
爱范儿
Google DeepMind News
Google DeepMind News
Stack Overflow Blog
Stack Overflow Blog
云风的 BLOG
云风的 BLOG
Engineering at Meta
Engineering at Meta
aimingoo的专栏
aimingoo的专栏
V
Visual Studio Blog
M
MIT News - Artificial intelligence
Hugging Face - Blog
Hugging Face - Blog

博客园_首页

No Articles

SolonCode v2026.5.21 released, Web can view projects, IM can find team members
带刺的坐椅 · 2026-05-21 · via 博客园_首页

1. About SolonCode (Terminal Coding Intelligent Agent)

SolonCode is an enterprise-level terminal coding intelligent agent developed by Hangzhou Wuye Technology Co., Ltd. It is a fully Chinese-driven digital employee—capable of understanding requirements autonomously, planning steps autonomously, and writing code autonomously. It doesn't matter what model or platform you use; just open the terminal to start working.

Core Differentiation: SolonCode vs Claude Code

Dimension SolonCode Claude Code
Language Environment Full Chinese guidance word driven, more in line with domestic development habits English为主
Openness 100% Open source, based on Java, flexible expansion Proprietary commercial product
Model dependency Unbound supplier, supports free configuration of mainstream large models Bound to Anthropic, supports configuration of other models
Interactive interface CLI / WEB / IM Robot CLI / APP

Open source project address:

2、Fast Installation and Getting Started

SolonCode has an extremely lightweight size (about 32MB) and supports mainstream operating systems with excellent Java environment compatibility.

System Requirements:

  • Java 8+ (supports Java 8 to Java 26 environments)
  • Supported on macOS, Linux, Windows

Installation or Update Commands (the package is automatically downloaded, about 32MB):

# Mac / Linux:
curl -fsSL https://solon.noear.org/soloncode/setup.sh | bash

# Windows (PowerShell):
irm https://solon.noear.org/soloncode/setup.ps1 | iex

Fast Installation and Update Demonstration:

3、Highlights of this Update

Since the release of v2026.5.13, SolonCode has undergone nearly 10 intensive iterations. This update revolves around three key terms:Web experienceIM connectionControl.

Web端,终于能“看”项目了

This update has added project browsing capabilities for the Web Interaction mode—left-side file tree provides an overview, keyword search locates results in seconds, and the right-side panel can be easily dragged and adjusted. No need to open an IDE, you can quickly browse code and copy snippets with one click in the browser. Additionally, a new Git Diff view has been added, making changes clear at a glance.

Now, usesoloncode web is launched, the browser becomes a lightweight AI IDE.

drag the AI coding assistant into the group chat

SolonCode is now integrated with WeChat, Lark, and DingTalk three major IM platforms. Scan to bind, and then mention it in the group or chat privately to directly issue coding and operations tasks. Lark uses WebSocket + Protobuf binary protocol, DingTalk is based on Stream long connection—messages are delivered and responded to instantly, just like real colleagues.

make AI a member of the team, no need to install new tools, find it anywhere.

more controllable conversations, more automated tasks

the new /rewind command supports rolling back the last N rounds of conversation—AI went off track? One-click to bring it back./loop Command adds Cron expressions, allowing precise scheduling of repetitive tasks to the second level (e.g., check deployment status every 5 minutes). When the Web input box is empty, the up and down arrow keys can directly select historical tasks, avoiding manual re-entry for repeated operations.

/loop cron:'0 */5 * * * ?' check status

Conversations have a "do-over" feature, tasks have timers.

There are also some noteworthy improvements

  • Performance : Context compression algorithm optimization, ReActAgent reasoning engine refactored into a more concise architecture, fixed multiple deadlock issues - overall more stable and faster
  • Mental memory : Persistent storage changed to Markdown files, the more you use it, the better it understands you, and you can edit it directly
  • @agent command : Specify sub-agents directly in the conversation to handle tasks, making division of labor more flexible
  • Skill tips:Web and CLI two interactive portals support skill auto-suggestion, reducing the learning curve.

4, Community and Feedback

Welcome to use and provide valuable suggestions! If you encounter any problems, you can provide feedback through the following ways: