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

What Is Easel

Easel is an AI social media content workbench jointly open-sourced by Zhejiang University and Peking University. The platform is based on OpenClaw Agent framework, using a single agent to connect the entire social media operations chain, from trend discovery, topic planning, and content creation to multi-platform publishing and data review. Through an account profile, it continuously remembers the account's positioning, style, and audience preferences. Together with 112 executable Skills, the AI understands you better the more you use it, becoming a content partner that truly "evolves."

Easel's Main Features

• Account Profile: Establish account context through positioning, style, audience, and red lines, with one profile used continuously across platforms.
• Trend Radar: Aggregate real-time trending lists from multiple platforms such as Weibo, Douyin, and Zhihu, and filter topics suitable for the account.
• Content Planning: Turn trends into topic matrices, titles, scripts, and schedules, and write them into the content calendar for unified management.
• Content Creation: Generate copy, posters, knowledge cards, voiceovers, AI videos, short dramas, and other publishable assets in multiple formats.
• Multi-Platform Publishing: Automatically adapt master content to each platform's format, and after quality checks, publish directly to logged-in accounts.
• Data Attribution: Collect playback, interaction, and comment data, and feed effective experience back into the profile, forming a loop that becomes more accurate the more it is used.
• Project Archiving: Save all outputs and metadata by project, making it easy to modify, retry, and reuse later.

Follow WeChat and reply "Open Source" to join theAI Open Source Project Exchange Group

How to Use Easel

• Environment Preparation: Ensure the system is Linux/macOS, and install Python 3.10+, Node.js 22.19+, and git.
• Clone the Repository: Run git clone git@github.com:ZJU-REAL/Easel.git to pull the code locally.
• Configure Keys: Copy .env.example as .env, and enter the LLM API Key (such as Anthropic).
• One-Click Installation: Run bash setup.sh to automatically complete the installation of Python dependencies, Node.js, and the OpenClaw framework.
• Environment Check: Run easel doctor and easel ping to confirm that the environment and gateway connection are normal.
• Install Extensions: If audio/video and publishing capabilities are needed, run pip install -e ".[media]" and playwright install chromium。
• Create a Profile: Copy the profiles/_template template or configure the account positioning, style, and audience on the Web side.
• Start the Workbench: Run easel web, and open http://localhost:7860 in the browser to enter the interface.
• Discover Trends: In "Trend Radar," aggregate trending lists from platforms such as Weibo and Douyin, and filter topics suitable for the account.
• Plan and Schedule: Use the "Content Calendar" to plan topics, titles, and scripts, forming an executable content matrix.
• Create Content: Through conversation or easel skill commands, let the Agent generate copy, cards, videos, and other assets.
• Check and Publish: In the "Publishing Center," complete sensitive-word/copyright checks and multi-platform format adaptation, then publish with one click.

Easel's Core Advantages

• Full-Chain Integration: A single Agent strings hot-topic discovery, topic planning, content creation, multi-platform publishing, and data review into a complete workflow.
• Profile-Driven Evolution: Through account profiles, it continuously remembers positioning, style, and audience, so each output fits the real account tone better than the last.
• Skills That Truly Execute: All 112 Skills are runnable scripts, not just Prompt templates, and can directly produce publishable image, text, audio, and video content.
• One Asset, Many Uses: The same master content automatically adapts to the format, tone, and aspect-ratio requirements of different platforms such as Xiaohongshu, Douyin, Zhihu, and Bilibili.
• Project-Based Archiving: All outputs are saved by project in outputs/, so edits, retries, and traceability never get scattered across chat logs.
• Closed-Loop Publishing Attribution: Supports real login and publishing on six major platforms, with publishing data automatically flowing back and accumulating, allowing account profiles to keep evolving.

Easel's Project Addresses

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

Comparison of Easel with Similar Competitors

Easel's Application Scenarios

• Daily operations for individual self-media bloggers: After the blogger configures the account profile, Easel automatically scrapes trending topics and generates image-text or video scripts tailored for Xiaohongshu/Douyin.
• Multi-platform matrix operations for brands/enterprises: A brand maintains accounts on multiple platforms. Easel automatically generates platform-adapted versions from a single master piece of content, unifying tone and enabling batch publishing, reducing repetitive cross-platform work.
• Batch content production for MCN agencies: An agency creates independent profiles for multiple influencer accounts under its management, using 112 Skills to batch-produce paper interpretations, spoken-word scripts, short videos, etc., archived by project for easy reuse and iteration.
• Content transformation for academic/knowledge accounts: A researcher inputs papers or professional knowledge into Easel. The Agent automatically generates knowledge cards, popular-science image-text posts, and video interpretations, transforming academic results into social-media-friendly content.