The recently launched general AI agent Manus has quickly sparked heated discussions in the Chinese AI community, with invitation codes becoming highly sought after. Many users have flocked to Discord, waiting for Tao Cheung to release new invites.

Manus’ official website provides 40 case studies demonstrating its capabilities across various domains, including travel planning, stock analysis, course development, insurance policy comparison, B2B supplier sourcing, financial report analysis, company list compilation, online store operations analysis, event explanation diagrams, candidate interview scheduling, lead generation, and press conference scripting.

From the official use cases, it is evident that Manus’ user interface design bears a strong resemblance to Devin, which made waves last year. Both feature modules like Browser, Shell, Editor, and Planner.

Introduction to Devin

Devin, developed by Cognition AI, is positioned as an autonomous AI software engineer capable of independently handling complex software development tasks, from coding to deployment. Its core strength lies in full-process task automation—users only need to issue instructions and later review progress and provide feedback. This sets Devin apart from tools that require constant monitoring and frequent interactions.

Currently, Manus remains in a closed beta phase, with no public testing available, though internal test results and case demonstrations are already circulating online. Meanwhile, Devin does not cater to individual users and offers a team version priced at $500 per month.

Tesla Stock Investment Analysis Case Study

To evaluate how Manus and Devin perform in practical applications, we conducted a comparative test using the same prompt in a Tesla stock analysis case study. This test assesses the AI agents’ ability to retrieve historical stock data, perform technical analysis, and integrate market trends and news events to generate actionable investment recommendations. Additionally, the AI should be able to autonomously develop an application to publish the analysis report as a publicly accessible website.

We used the exact same prompt featured in Manus’ official case study, without any modifications to the interaction process or deployment flow.

Initial Prompt Used for the Test

I'd like a thorough analysis of Tesla stock, including:

Summary: Company overview, key metrics, performance data and investment recommendations Financial Data: Revenue trends, profit margins, balance sheet and cash flow analysis Market Sentiment: Analyst ratings, sentiment indicators and news impact Technical Analysis: Price trends, technical indicators and support/resistance levels Compare Assets: Market share and financial metrics vs. key competitors Value Investor: Intrinsic value, growth potential and risk factors Investment Thesis: SWOT analysis and recommendations for different investor types

Execution Process and Interface Comparison

The Manus interface demonstrates a structured approach to task execution. Like Devin, Manus creates a To-Do plan, but instead of integrating it directly into the UI like Devin, it generates a Markdown file to track task execution. Notably, Devin also uses a todo.md file in real-world implementations, such as within Impa’s daily operations, to document plan execution status.

Both AI agents provided detailed reports and visual charts in their final outputs.

Report Presentation Comparison

Upon deployment, the final reports on the websites differed in presentation:

Manus

Devin

• Manus’ report follows a professional investment analysis layout, placing the Executive Summary and Recommendations at the top, followed by supporting data visualizations and financial metrics.

• Devin’s report, in contrast, appears less refined in terms of UI and data visualization, with slightly less polished chart selections and storytelling structure.

Given that both AI agents possess Knowledge Base/Memory features, there is a possibility that additional knowledge sources contributed to the differences in presentation.

Optimized Prompt via Devin’s Built-in Enhancement Feature

A notable advantage of Devin is its ability to optimize the initial prompt. When a user inputs a request, Devin suggests refinements, prompting them to specify data sources, APIs, and tools while breaking down the task into more manageable sub-tasks.

Similarly, ChatGPT’s Deep Research mode has introduced an iterative confirmation mechanism, ensuring precision in task execution by repeatedly clarifying details before proceeding.

The revised prompt after Devin’s optimization:

Please help analyze Tesla (TSLA) stock and create a comprehensive investment analysis report.

Required analysis components:

• Company overview including key metrics and recent performance

• Financial analysis of revenue trends, margins, balance sheet and cash flows

• Market sentiment analysis covering analyst ratings and news impact

• Technical analysis of price trends and key support/resistance levels

• Competitive analysis comparing market share and financials vs key competitors

• Valuation analysis including intrinsic value estimates and growth potential

• SWOT analysis and investment recommendations for different investor types

Please compile the analysis into a well-structured report with clear sections. No need to test locally - just create the analysis report based on available public data.

Website Implementation and Technical Stack Differences

Manus

Devin

• Manus and Devin (Same Prompt): Both used lightweight Chart.js for data visualization and vanilla HTML/CSS/JavaScript for rendering.

• Devin (Optimized Prompt): The enhanced version leveraged React, Vite, Tailwind CSS, and Recharts, which are preferred by React developers for greater customization and interactive charts.

Manus vs. Devin: Key Differentiators

Manus: Intelligent Analysis-Oriented Product Design

Manus behaves more like a highly efficient AI analyst, focusing on task feedback and tangible output. When given a clear directive, Manus delivers well-structured, visually appealing analysis reports, catering to users who require instant insights and data-driven recommendations.

Devin: Engineering-Oriented Product Design

Devin, in contrast, functions more like a junior developer within a team, emphasizing engineering workflow optimization. Beyond solving the task at hand, Devin integrates with team collaboration tools like Slack, GitHub (automatic PRs), Secret storage, and Cursor integration, making it more suitable for software development projects requiring long-term maintenance and scalability.

Final Thoughts: The Future of AI Agents

As AI agents continue to evolve, the boundaries between instant insights and engineering-centric solutions are gradually merging. Future models will likely harmonize these strengths, balancing on-demand analysis with long-term software development capabilities.

Both Manus and Devin offer unique advantages, and choosing the right AI agent depends on specific use cases. Over time, large models will integrate these capabilities, delivering even more versatile, intelligent, and productive AI experiences.

References

• Manus AI

• Manus Tesla Stock Analysis Demo

• Manus Website Report

• Devin AI