Devtoolwatch BoardDSH-IM review
DSH-IM is a DeepSeek Harness plugin that connects IM bots such as Feishu, WeChat, Telegram, Slack and WhatsApp to a local Harness.
TL;DR
DSH-IM is a plugin for DeepSeek Harness (DSH) that connects chat bots on many IM platforms to a local Harness. Users bind a bot by QR code, App Manifest or existing credentials. DSH-IM fits developers who want to run Harness tasks from chat apps. Each channel supports multiple bots, and each bot keeps its own workspace, model and session bindings.
- 3kGitHub stars
- +588new stars, week of Oct 25
- ~45.8kdownloads a week, +117% vs 4-week avg
- #25in Coding agents this week
DSH-IM at a glance
- Category
- Coding agents
- Maintained by
- xiemanR (individual, @xmanrui on GitHub)
- License
- MIT
- Language
- JavaScript, Shell
- Repository
- github.com/xmanrui/dsh-im
- npm package
- @xmanrui/dsh-im 5.6.0
- Latest release
- v5.6.0 on GitHub, Oct 24, 2026
- First published
- Aug 14, 2026
What DSH-IM does
- DSH-IM connects bots on Feishu, WeChat, DingTalk, WeCom, QQ, Slack, Telegram, Discord, WhatsApp, iMessage and Matrix (experimental) to DeepSeek Harness.
- DSH-IM supports multiple bots per channel, and each bot has its own workspace, model, Agent Preset and session bindings.
- DSH-IM accepts JPEG, PNG, WebP and GIF images as input and can send files the Harness can read back as native attachments.
- DSH-IM can archive received group messages and attachments locally on six channels, keeping them for 14 days by default.
- DSH-IM provides chat commands such as /new, /model, /workspace, /stop, /steer and /compact to control Harness sessions.
- DSH-IM supports remote approvals from chat by replying yes or no, plus proactive message delivery to saved targets.
- DSH-IM resends final text results after a model reply timeout once the original task completes.
How it works
DSH-IM installs as a plugin into DSH Desktop or DSH Web and is managed under Settings → IM Bots. On a local host, DSH-IM reuses the host's internal services through Typert Gateway to call Session Controller and Workspace Controller. No Harness address is needed. Secrets and tokens are sent only to the local Harness Host and kept in protected credential storage. DSH-IM can also connect the local Harness outward to a public AI Office.
How to install and start using DSH-IM
DSH Desktop (built-in terminal)
dsh plugin --profile desktop add -w @xmanrui/dsh-imDSH Web (terminal that started Harness)
dsh plugin --profile web add -w @xmanrui/dsh-imLatest code from GitHub
npx -y github:xmanrui/dsh-im installRequirements
- DeepSeek Harness (DSH Desktop or DSH Web) installed locally.
- Node.js 22.19 or later.
- Node.js 22.21 or later when reaching the Telegram Bot API through a proxy.
- A bot account or credentials for each IM platform, such as a Telegram Bot Token or Slack Bot and App Tokens.
First steps
- Install DSH-IM with the command that matches DSH Desktop or DSH Web, using the full package name @xmanrui/dsh-im.
- Restart DSH Desktop, or restart dsh web and refresh the browser.
- Open Settings → IM Bots and pick a channel.
- Follow the built-in guide to scan a QR code or enter bot credentials.
Who DSH-IM is for
- DeepSeek Harness users who want to run agent tasks from chat apps.
- Developers managing several bots with separate workspaces and models.
Limitations to know
- DSH-IM works only as a plugin for DeepSeek Harness, not as a standalone bot.
- DSH-IM's Matrix encryption is experimental and for non-sensitive testing only. It does not provide full end-to-end confidentiality.
- DSH-IM supports only text private chats on iMessage, with no image input or file return.
- DSH-IM's two-way session sync supports only private-chat text on the current host. Groups, topics, threads and remote harnessBaseUrl are excluded.
Is DSH-IM growing? Usage data
In the week of Oct 25–31, 2026, DSH-IM ranked #42 on the Devtoolwatch Board and #25 in Coding agents, with a momentum score of 78 out of 100. Its main package had about 45.8k downloads that week, +117% against its own 4-week average.
Show the weekly numbers
| Week of | New stars |
|---|---|
| Aug 9, 2026 | 6 |
| Aug 16, 2026 | 552 |
| Aug 23, 2026 | 405 |
| Aug 30, 2026 | 133 |
| Sep 6, 2026 | 155 |
| Sep 13, 2026 | 128 |
| Sep 20, 2026 | 107 |
| Sep 27, 2026 | 84 |
| Oct 4, 2026 | 196 |
| Oct 11, 2026 | 330 |
| Oct 18, 2026 | 341 |
| Oct 25, 2026 | 588 |
The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.
DSH-IM alternatives
Other AI coding agents on the Board in the week of Oct 25–31, 2026, by momentum:
| Rank | Tool | Stars | Downloads/wk |
|---|---|---|---|
| #1 | CindyCindy is an open-source AI agent app for desktop and mobile that combines Claude Code, Codex and Pi across your connected devices. | 8.8k | – |
| #2 | OpenMausBotOpenMausBot is an open-source desktop chat app where each contact is a local Claude, Codex, or Grok CLI agent with its own computer. | 9.6k | – |
| #3 | OrcaOrca is a desktop agent development environment for running Claude Code, Codex, and other CLI coding agents in parallel, each in its own… | 111.7k | – |
| #4 | OpenBotOpenBot is an open-source, self-hosted template for AI coworkers that each get their own browser, files and granted tools, with every… | 9.1k | – |
| #5 | Agent ReachAgent Reach is a Python CLI that lets AI agents search and read platforms such as Twitter, Reddit, YouTube, GitHub, Bilibili and… | 107.2k | – |
| #6 | Skill RecorderSkill Recorder is a desktop app that records your on-screen work session and turns it into a reusable agent Skill or Automation. | 5.6k | – |
Recent DSH-IM releases
DSH-IM FAQ
Is DSH-IM free and open source?
Yes. DSH-IM's source code is on GitHub (xmanrui/dsh-im) under the MIT license, so you can use it for free.
How do I install DSH-IM?
Run dsh plugin --profile desktop add -w @xmanrui/dsh-im. It needs Node.js >=22.19. The install section above has the other options from the project's README.
Is DSH-IM still maintained?
Yes. The xmanrui/dsh-im repository was last updated on Oct 25, 2026, and the latest release, v5.6.0 on GitHub, came out on Oct 24, 2026.
How popular is DSH-IM?
DSH-IM has 3k GitHub stars as of Nov 1, 2026. It gained 588 new stars in the week of Oct 25–31, 2026, against a 12-week average of 252 a week. Its main package had about 45.8k downloads that week, +117% against its own 4-week average. That made it #25 among AI coding agents on the Devtoolwatch Board.
What are the best alternatives to DSH-IM?
Other AI coding agents on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: Cindy (#1), OpenMausBot (#2), Orca (#3), OpenBot (#4), Agent Reach (#5). The alternatives table above has their numbers.
Which chat platforms does DSH-IM support?
DSH-IM supports Feishu, WeChat, DingTalk, WeCom bots and apps, QQ, Slack, Telegram, Discord, WhatsApp, iMessage on macOS, and Matrix as an experimental channel.
Where does DSH-IM store bot secrets?
DSH-IM sends secrets and tokens only to the local Harness Host, which writes them to protected credential storage. Status endpoints and bot lists do not return these credentials.
Can DSH-IM bots use different models?
Yes. In DSH-IM, each bot can select its own model and reasoning level. A change applies only to new sessions, so send /new before the next message to use it.
What happens to new messages while a DSH-IM task is running?
DSH-IM queues new messages by default until the current task ends. A per-bot steer mode instead adds plain text as supplementary instructions to the running task. /steer works in both modes.
Does DSH-IM need a proxy configuration?
DSH-IM needs a proxy only if the machine cannot reach a platform directly. For Feishu, set HTTPS_PROXY before starting dsh web. For Telegram, use Node.js 22.21 or later with NODE_USE_ENV_PROXY=1, then restart the Host.
Sources
DSH-IM README badge
Maintainers can show the tool's weekly Board rank in their README:
[](https://index.devtoolwatch.com/tools/xmanrui-dsh-im/)How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].