Easel – An AI-Powered Social Media Content Workbench Open-Sourced by Zhejiang University and Peking University

Easel – An AI-Powered Social Media Content Workbench Open-Sourced by Zhejiang University and Peking University

2026-09-03 10:00

Introduction: Powered by OpenClaw Agent, 112 executable Skills drive the full-stack workflow. What is Easel? Easel is an open-source AI social media content workspace jointly developed by Zhejiang University and Peking University.

What is Easel?

Easel is an open-source AI social media content workspace co-developed by Zhejiang University and Peking University. Built on the OpenClaw Agent framework, the platform leverages a single intelligent agent to unify the entire social media operation pipeline—from trend discovery and topic planning to content creation, multi-platform publishing, and data retrospective analysis. By continuously maintaining a user profile that captures account positioning, stylistic identity, and audience preferences, combined with 112 executable Skills, Easel evolves with usage, becoming a truly "self-evolving" content collaborator.

Easel

Core Features of Easel

  • Account Profiling: Establishes contextual awareness through positioning, style, audience, and compliance boundaries—enabling cross-platform consistency via a single, persistent profile.
  • Trend Radar: Aggregates real-time trending lists from Weibo, Douyin, Zhihu, and other platforms to identify topics aligned with your account’s niche.
  • Content Planning: Transforms trends into structured topic matrices, titles, scripts, and scheduling plans—organized in a unified content calendar.
  • Content Generation: Produces diverse publishable assets including copywriting, posters, knowledge cards, voiceovers, AI-generated videos, and short-form dramas.
  • Multi-Platform Publishing: Automatically adapts master content to each platform’s format requirements, performs quality checks, then publishes directly to logged-in accounts.
  • Data Attribution: Collects playback, engagement, and comment data, feeding validated insights back into the profile to create a continuously improving feedback loop.
  • Project Archiving: Saves all outputs and metadata per project within outputs/, enabling future revisions, re-runs, and reuse without scattered chat logs.

Dig GitHub

Follow WeChat and reply “open source” to join the AI Open Source Project Community

How to Use Easel

  • Environment Setup: Ensure Linux or macOS system; install Python 3.10+, Node.js 22.19+, and git.
  • Clone Repository: Execute git clone git@github.com:ZJU-REAL/Easel.git to pull code locally.
  • Configure Keys: Copy .env.example to .env, then fill in LLM API Key (e.g., Anthropic).
  • One-Click Installation: Run bash setup.sh to automatically install Python dependencies, Node.js, and the OpenClaw framework.
  • Environment Validation: Execute easel doctor and easel ping to confirm proper environment setup and gateway connectivity.
  • Install Extensions: For audio/video and publishing capabilities, run pip install -e ".[media]" and playwright install chromium.
  • Create Profile: Duplicate the profiles/_template folder or configure account positioning, tone, and audience via the Web UI.
  • Launch Workspace: Run easel web and access the interface at http://localhost:7860 in your browser.
  • Discover Trends: In the “Trend Radar,” aggregate hotlists from Weibo, Douyin, etc., and filter relevant topics for your account.
  • Plan Schedule: Use the “Content Calendar” to schedule topics, headlines, and scripts, forming an executable content matrix.
  • Generate Content: Leverage conversational prompts or easel skill commands to instruct the Agent to produce copy, cards, videos, and more.
  • Review & Publish: Complete sensitive word checks, copyright validation, and cross-platform formatting in the “Publish Center,” then one-click publish.

Core Advantages of Easel

  • Full-Stack Integration: A single Agent seamlessly connects trend discovery, topic planning, content generation, multi-platform publishing, and post-pub analytics into a cohesive workflow.
  • Profile-Driven Evolution: Persistent account profiling remembers positioning, style, and audience behavior—each output becomes increasingly aligned with the actual account identity over time.
  • True Skill Execution: The 112 Skills are executable scripts—not just prompt templates—capable of directly generating publishable text, images, audio, and video content.
  • One Source, Multiple Outputs: A single master asset is automatically adapted to meet the unique format, tone, and aspect ratio demands of Xiaohongshu, Douyin, Zhihu, Bilibili, and WeChat Video Channels.
  • Project-Based Archiving: All outputs are saved per project under outputs/, ensuring traceability, reusability, and version control—free from fragmented chat histories.
  • Closed-Loop Attribution: Supports real login and publishing on six major platforms; post-publishing data auto-flows back to refine the profile, enabling continuous evolution.

Project Links

  • Official Website: https://zju-real.github.io/Easel/
  • GitHub Repository: https://github.com/ZJU-REAL/Easel

Competitive Comparison

Comparison Dimension Easel Postiz
Development Team Zhejiang University REAL Lab × Peking University OpenDCAI Lab GitroomHQ (Overseas Team)
GitHub Stars 221+ 35,000+
Open Source License Apache 2.0 AGPL-3.0
Core Positioning AI Agent-powered content workspace—unifying the full workflow from discovery to creation, publishing, and attribution Open-source social media management and scheduling platform, excelling in scheduling, publishing, and team collaboration
AI Depth Built on OpenClaw Agent framework with 112 executable Skills autonomously handling trend detection, topic planning, and multi-format content generation Includes AI-assisted title, copy, and tag generation—but core content still relies on manual user input
Content Production Agent-driven: From trend radar to final published assets (text, image, audio, video), with project-based archiving in outputs/ Users upload or paste content; AI only provides polishing and optimization suggestions
Account Profiling Six-dimensional profile (positioning, style, audience, platforms, red lines, long-term memory); data auto-recycled for continuous evolution Basic brand configuration (logo, color, font); lacks persistent learning mechanism
Platform Coverage Focuses on Chinese social platforms: Xiaohongshu, Douyin, Kuaishou, Bilibili, Zhihu, WeChat Video Channel Covers 25+ international platforms: Instagram, TikTok, X, LinkedIn, Facebook, etc.
Publishing Method Browser automation for real publishing with preview and quality gate (sensitive words, copyright, SEO checks) Official API-driven publishing—stable and reliable, supports scheduled posting
Data Feedback Loop Automatically collects playback, engagement, and comment data, feeds it back into the profile to form a self-improving closed loop Provides basic engagement analytics dashboard—no automatic attribution or profile learning

Use Cases of Easel

  • Daily Operations for Individual Content Creators: After configuring their account profile, creators can leverage Easel to automatically discover trending topics and generate optimized scripts tailored for Xiaohongshu or Douyin.
  • Multi-Platform Matrix Management for Brands/Enterprises: Brands managing multiple accounts use Easel to generate platform-specific versions from a single master asset, ensuring consistent tone and enabling bulk publishing—reducing redundant cross-platform work.
  • Bulk Content Production for MCNs: MCNs establish independent profiles for multiple influencer accounts and use the 112 Skills to mass-produce research summaries, spoken scripts, short videos, etc.—archived per project for reuse and iteration.
  • Content Transformation for Academic/Knowledge Accounts: Researchers input papers or expertise into Easel, where the Agent generates knowledge cards, science communication graphics, and video explainers—transforming academic findings into social-friendly content.

Source: AI Tool Kit

#AI Tools

Disclaimer: Contains third-party opinions, does not constitute financial advice

Share To
X
Telegram
WeChat
QQ
Link
Recommended Reading

Hy4 Preview Lightweight Version – Tencent Hunyuan's Open-Source Hy4 Quantized Compression Model

1 day ago
Hy4 Preview Lightweight Version – Tencent Hunyuan's Open-Source Hy4 Quantized Compression Model

Claude Fable 5.1 – Anthropic's Latest Flagship Large Model

1 day ago
Claude Fable 5.1 – Anthropic's Latest Flagship Large Model

Open-source Retool is here, drag-and-drop backend with 60+ components

15 days ago
Open-source Retool is here, drag-and-drop backend with 60+ components

Claude Opus 5 is Here: Half the Price Approaching Fable 5

07-25
Claude Opus 5 is Here: Half the Price Approaching Fable 5

Tencent Open-Sources 260 Real-World Tasks to Test AI Workforce Capabilities

07-25
Tencent Open-Sources 260 Real-World Tasks to Test AI Workforce Capabilities

Meituan Open-Sources AI Exam Platform: 813 Minecraft Tasks

07-25
Meituan Open-Sources AI Exam Platform: 813 Minecraft Tasks

Microsoft's Image Model Integrated into PPT, GPU Costs Reduced by 84%

07-24
Microsoft's Image Model Integrated into PPT, GPU Costs Reduced by 84%