X

AgentScope

Information

## AgentScope The core framework for agent development, features a highly modular design and asynchronous architecture, enabling flexible tool invocation, real-time intervention control, and intelligent context management. ## Introduction With the rapid advancement of Large Language Models (LLMs), significant progress has been made in multi-agent applications. However, the complexities in coordinating agents’ cooperation and LLMs’ erratic performance pose notable challenges in developing robust and efficient multi-agent applications. To tackle these challenges, we propose AgentScope, a developer-centric multi-agent platform with message exchange as its core communication mechanism. The abundant syntactic tools, built-in agents and service functions, user-friendly interfaces for application demonstration and utility monitor, zero-code programming workstation, and automatic prompt tuning mechanism significantly lower the barriers to both development and deployment. Towards robust and flexible multi-agent application, AgentScope provides both built-in and customizable fault tolerance mechanisms. At the same time, it is also armed with system-level support for managing and utilizing multi-modal data, tools, and external knowledge. Additionally, we design an actor-based distribution framework, enabling easy conversion between local and distributed deployments and automatic parallel optimization without extra effort. With these features, AgentScope empowers developers to build applications that fully realize the potential of intelligent agents. We have released AgentScope at https://github.com/modelscope/agentscope, and hope AgentScope invites wider participation and innovation in this fast-moving field.

Prompts

1

Navigate to arxiv.org and take a screenshot of the latest papers

Reviews

Tags


  • rockingdingo 2025-09-05 14:16
    Interesting:4,Helpfulness:4,Token Consumption:2,Correctness:3
    Prompt: Navigate to arxiv.org and take a screenshot of the latest papers

    I just tested the browser agent in the release demo and asked the bot agent to navigate to arxiv.org find the latest papers and take a screenshot of the pages. The whole workflow requires a chain of tool call of navigate->click->screenshot. Overall, it succeeded in the first step and failed the remaining. The token consumption is high and context prompt usage is not ideal. The framework is still easy to understand though. Pros: 1. The agent identify the first tool and successfully navigate to the arxiv.org. This step is successful. Cons: 1. To find the latest paper, it should navigate to the new section of the website, the tool is correct but the parameters are not right. And all the sudden the workflow breaks because of the error An error occurred: Failed to get response from _ API: {"status_code": 400, "request_id": "23c1b468-db66-4109-ae9d-30a99b58591f", "code": "InvalidParameter", "message": " InternalError.Algo.InvalidParameter: Range of input length should be [1, 30720]", "output": null, "usage": null}. 2. Token Consumption Why would the agent send the whole webpage html code to the LLM? The result is not necessary for the screenshot tasks and it consumes too much tokens. And the error handling is not good. 3. The screenshot task is not finished.

Write Your Review

Detailed Ratings

ALL
Correctness
Helpfulness
Interesting
Upload Pictures and Videos

Name
Size
Type
Download
Last Modified

Upload Files

  • Community

Add Discussion

Upload Pictures and Videos