CowAgent
Self-hosted AI agent (formerly chatgpt-on-wechat) that answers in WeChat, WeCom, Feishu, DingTalk, Slack and a web console.
Quick answer
CowAgent is a self-hosted AI agent (formerly chatgpt-on-wechat) that answers in WeChat, WeCom, Feishu, DingTalk, Slack and a web console, listed under AI app & agent builders. A typical small-business use: put an AI assistant into the WeCom, Feishu or DingTalk groups your team already uses. The project publishes no official RAM figure, so we do not estimate a monthly server cost. License: MIT, which allows internal business use.
Last updated · facts verified against official docs on · sources
Activity data comes from GitHub and was last refreshed Oct 30, 2026.
What is CowAgent?
CowAgent, previously chatgpt-on-wechat, is an AI assistant you run on your own computer or server. One instance serves a web console and several chat channels at once, including WeChat, WeCom, WeChat Customer Service and Official Accounts, Feishu/Lark, DingTalk, QQ, Telegram and Slack. It plans tasks with built-in tools (files, terminal, browser, web search, scheduler) and MCP, keeps long-term memory and a knowledge base, and works with most major model providers, configured from the web console.
Where it fits in a 10–200 person business
- Put an AI assistant into the WeCom, Feishu or DingTalk groups your team already uses
- Answer customers in WeChat Customer Service or an Official Account with your own knowledge base
- Run scheduled reports or reminders that the agent posts back to a chat channel
These use cases are our editorial suggestions, not claims made by the project.
Can a business use CowAgent? (license)
| License | MIT |
|---|---|
| Type | Permissive: MIT / Apache-2.0: use, modify and sell with few obligations. |
| Internal business use | Allowed under the license text |
| What to watch | Permissive. Internal and commercial use allowed; keep the copyright notice. |
This is a plain-English summary of the license text, not legal advice. Third-party models you run with it have their own licenses.
Data & compliance: does CowAgent phone home?
| Telemetry (usage data sent to the vendor) | Not documented |
|---|---|
| Other calls home by default (updates, marketplaces, default cloud services) | Not documented |
| Single sign-on (OIDC / SAML / LDAP) | Not documented |
| Audit log | Not documented |
| Role-based access control (RBAC) | Not documented |
| Data residency / GDPR notes from the project | Not documented |
Checked against the project's own docs and repository on . "Not documented" means we could not find the answer in official sources; it does not mean the feature is missing or that no data is sent. Third-party services you connect (model APIs, vector databases, search engines) have their own data policies. Compare all apps.
What hardware does CowAgent need?
The project does not publish a minimum CPU/RAM figure, so we leave the official value blank rather than guess.
GPU: No. It calls the model providers you configure.
No official CPU, RAM or disk figure is published. The script install needs Python 3.7-3.13 (3.9 recommended).
How much does it cost to host CowAgent per month?
The project publishes no RAM figure we could verify, so we do not estimate a cost.
How to deploy CowAgent
Method: Docker Compose (official compose file)
curl -O https://cdn.link-ai.tech/code/cow/docker-compose.yml
docker compose up -d
# Web console on http://localhost:9899From the official instructions: github.com/zhayujie/CowAgent#readme. Always check the source for the current version before you run it.
Known gotchas when self-hosting CowAgent
- On a server, the console is only reachable after setting web_host to 0.0.0.0 in config.json; set web_password at the same time to protect it, and restrict port 9899 to specific IPs. source
- The agent has access to the host operating system through tools such as the terminal and file writes, so the maintainers say to deploy it only in trusted environments. source
- Agent mode uses substantially more tokens than plain chat; pick models with cost in mind. source
Alternatives in AI app & agent builders
- AutoGPT Platform Visual builder for long-running AI agents and workflows, self-hostable as a single container. Compare AutoGPT Platform vs CowAgent
- Dify Visual platform for building LLM apps, RAG chatbots, agents and workflows. Compare CowAgent vs Dify
- Langflow Visual builder for AI agents and RAG flows that you can deploy as APIs or MCP servers. Compare CowAgent vs Langflow
- Mem0 Long-term memory layer for AI assistants and agents, with a self-hosted REST server and dashboard. Compare CowAgent vs Mem0
- Flowise Drag-and-drop builder for LLM agents and chatflows (archived August 2026).
- AstrBot Multi-platform AI chatbot and agent framework for QQ, WeCom, Feishu, DingTalk, Official Accounts, Telegram and Slack. Compare AstrBot vs CowAgent
- Sim Visual builder for AI agent workflows, with knowledge bases, integrations and an API. Compare CowAgent vs Sim
- Coze Studio Open-source edition of ByteDance's Coze agent-building platform. Compare CowAgent vs Coze Studio
CowAgent FAQ
How much RAM does CowAgent need?
CowAgent does not publish an official RAM figure, so we do not state one. The script install needs Python 3.7-3.13 (3.9 recommended).
Does CowAgent need a GPU?
No. It calls the model providers you configure.
How much does it cost to self-host CowAgent?
The software is free to download. The project publishes no official RAM figure, so we do not estimate a monthly server cost.
Is CowAgent free for commercial use?
CowAgent is licensed under MIT (Permissive). Internal business use is allowed under the license text. Permissive. Internal and commercial use allowed; keep the copyright notice. This is a summary, not legal advice.
What are common problems when self-hosting CowAgent?
On a server, the console is only reachable after setting web_host to 0.0.0.0 in config.json; set web_password at the same time to protect it, and restrict port 9899 to specific IPs. The agent has access to the host operating system through tools such as the terminal and file writes, so the maintainers say to deploy it only in trusted environments. Agent mode uses substantially more tokens than plain chat; pick models with cost in mind.
Is CowAgent still maintained?
Yes, it is actively developed. Last commit: Oct 30, 2026. Latest release: 2.2.2 (Oct 30, 2026). Checked Oct 30, 2026.
How do I install CowAgent?
The official method is Docker Compose (official compose file). The exact commands are in the "How to deploy" section above and in the official instructions: https://github.com/zhayujie/CowAgent#readme
What are self-hosted alternatives to CowAgent?
Other apps in AI app & agent builders: AutoGPT Platform, Dify, Langflow, Mem0, Flowise, AstrBot, Sim, Coze Studio.
Sources
- GitHub repository: github.com/zhayujie/CowAgent
- Official documentation, Hardware requirements, Gotcha 1: docs.cowagent.ai/guide/quick-start
- License text: github.com/zhayujie/CowAgent/blob/master/LICENSE
- Deploy instructions, Gotcha 2, Gotcha 3: github.com/zhayujie/CowAgent#readme
- Hetzner Cloud pricing (cost estimate): www.hetzner.com/cloud/regular-performance
- GitHub stars, commits and releases: GitHub REST API, fetched
Facts on this page were checked against the project's own repository and docs on 2026-10-30. Spotted something out of date? Tell us.