Skip to main content

MCP Server & CLI for social media automation — let AI agents publish to TikTok, Instagram, YouTube, Douyin, XHS, Bilibili and 15+ platforms via Model Context Protocol

Project description

HuiMei 慧媒 — Chinese Social Media Automation CLI & MCP Server

PyPI version Python License: MIT

HuiMei provides a CLI and MCP interface for publishing to supported Chinese social-media platforms through cloud or local browser execution.

Why HuiMei?

  • 🌏 Multi-platform workflow — One command surface for the platforms and content types reported by huimei platforms.
  • 🔀 CLI + MCP dual interface — Humans use the terminal (huimei publish), AI agents use the MCP protocol. Same engine, two entry points.
  • 🔒 Cloud, local, and auto execution — Local mode drives the user's Chrome through a managed Runtime Host; cloud mode runs the same signed automation bundle on the service environment.
  • 🔑 Protected credentials and runtime code — Sensitive payloads use HTTPS/WSS plus application-layer encryption. Local tasks receive short-lived, device-bound encrypted credential envelopes, and automation code is delivered as a signed encrypted Runtime Bundle.
  • 🛡️ Upload-safe preflight — CLI and MCP require a target-bound server Runtime/Catalog preflight before uploading media or creating a publish task.

Supported Platforms

Platform ID Name Video Image
douyin 抖音 (Douyin)
xhs 小红书 (Xiaohongshu/RED)
bilibili B站 (Bilibili)
ks 快手 (Kuaishou)
weibo 微博 (Weibo)
tencent 视频号 (WeChat Channels)
tk TikTok
toutiao 头条 (Toutiao)
baijiahao 百家号 (Baijiahao)
weixingongzhonghao 微信公众号 (WeChat MP)
zhihu 知乎 (Zhihu)

Installation

Option 1: pipx (recommended)

pipx install huimei

Option 2: One-line install script

curl -fsSL https://huimei.smaroot.tech/install.sh | bash

Option 3: pip inside a virtual environment

python3 -m venv .venv
source .venv/bin/activate
python -m pip install huimei

Quick Start

# Log in to your account
huimei login

# List supported platforms
huimei platforms

# List linked social media accounts
huimei account list

# Actively refresh account status and related work data
huimei account check --account <id> --wait --json

# Synchronize, list and inspect work statistics
huimei works sync --platform zhihu --wait --json
huimei works list --json
huimei works detail <work-id> --json

# Publish using the stable account ID (recommended)
huimei publish -p zhihu --account-id <id> -v ./video.mp4 -t "Title" --json

# Check CLI and connection status
huimei status

# View or update the global publish mode saved on the server
huimei mode
huimei mode auto

# 客户端发布固定使用 headless;账号登录仍会打开有头浏览器
huimei browser show
huimei browser set --chrome-path "/path/to/chrome" --window-size 1440x900

# Machine-readable capabilities and asynchronous task status for AI Agents
huimei agent-context
huimei task-status <task-id> --json

MCP Server Integration

MCP (Model Context Protocol) allows AI agents to call external tools as if they were native functions. HuiMei ships a built-in MCP server so agents like Claude, Hermes, and GPT can manage your social media directly.

Claude Code / Claude Desktop

Add to your MCP config (~/.claude/claude_code_config.json or Claude Desktop settings):

{
  "mcpServers": {
    "huimei": {
      "command": "huimei-mcp-server",
      "args": []
    }
  }
}

Hermes Agent

Add to your Hermes config:

mcp_servers:
  huimei:
    command: huimei-mcp-server

Available MCP Tools

Tool Description
huimei_login Launch browser login flow for a platform
huimei_status Check CLI status and backend connectivity
huimei_logout Log out and clear local session
huimei_platforms List all supported platforms and capabilities
huimei_accounts List all linked social media accounts
huimei_account_check Refresh account status with batch semantics
huimei_works_sync Synchronize account and work data
huimei_works_list List synchronized works and statistics
huimei_works_detail Read one synchronized work
huimei_capabilities Read the authoritative runtime contract
huimei_diagnose Check local publish-device readiness
huimei_account_login_start Start interactive browser login asynchronously
huimei_local_task_status Query browser-login task status
huimei_publish Create an asynchronous publish task
huimei_publish_status Query the final platform publish status
huimei_set_mode Update the global publish mode

Architecture

CLI / MCP / Web
      │ HTTPS + WSS
      ▼
Huimei Platform ── task, account, device and status coordination
      │
      ├── cloud ──▶ Runtime Host ──▶ managed Python + encrypted Runtime Bundle
      │
      └── local ──▶ Local Worker ──▶ managed Python + encrypted Runtime Bundle ──▶ Chrome

The CLI contains the stable command, transport, Worker, and Runtime Host layers. Platform automation code and its dependency descriptor are versioned independently, so script updates do not normally require a CLI upgrade. Publishing is asynchronous; task creation is not treated as final platform success.

Use Cases

📹 Content Creators — Write once, publish everywhere. Record a video, add your caption, and distribute to all 11 platforms with a single command.

🤖 AI Agents — Let Claude or GPT manage your social media via MCP. Your agent can check account status, schedule posts, and publish content autonomously.

📊 Marketing Teams — Automate matrix account publishing across dozens of accounts and platforms. Integrate into your existing CI/CD or content pipeline.

Links

License

MIT © 2026 Beijing HuiGen ZhiYan Technology Co., Ltd (北京慧根智研科技有限责任公司)

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

huimei-0.6.59-py3-none-any.whl (246.6 kB view details)

Uploaded Python 3

File details

Details for the file huimei-0.6.59-py3-none-any.whl.

File metadata

  • Download URL: huimei-0.6.59-py3-none-any.whl
  • Upload date:
  • Size: 246.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.11.15

File hashes

Hashes for huimei-0.6.59-py3-none-any.whl
Algorithm Hash digest
SHA256 eafc414ad3440d5ae204929d3e1e41edd6a568e506918bafb0c4b4532b860f6a
MD5 27ee40525f5bffc723eeb810b245f89e
BLAKE2b-256 5fb7cd698774ace1047f206e17924eca2e41ae2f2b36bd0240c09bcd2c4d1b34

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page