Nowadays, developers can shift their attention from the pursuit of Manus invitation codes and turn to the vast world of the open source community. The CAMEL-AI team recently announced the launch of an innovative project called OWL (Optimized Workforce Learning). This multi-agent collaboration framework based on the CAMEL-AI framework has set off a revolutionary storm in the field of automated task processing with its outstanding performance and open concept.
OWL achieved remarkable results in the famous GAIA benchmark, topping the open source framework with an average score of 58.18. This achievement not only surpassed Huggingface's Open Deep Research, but also made people wonder about Manus' future. The CAMEL-AI team responded with Geek-style humor: "That's it? Let's start replicating in 0 days!"
In sharp contrast to Manus' mysterious internal beta and sky-high invitation code, OWL's biggest highlight is its fully open source feature. Developers can directly obtain code on GitHub, experience this powerful framework personally, and participate in the construction of OWL to jointly create a more powerful all-round open source agent. The project address has been published: https://github.com/camel-ai/owl.
OWL's intelligent performance is far beyond expectations, and its core lies in flexible and efficient multi-agent collaboration. Through dynamic agent interaction, OWL achieves more natural, efficient and robust task automation. It not only makes multiple AIs work together, but also focuses on the collaboration patterns and communication protocols between them.
OWL's powerful task automation capabilities are designed to promote the boundaries of task automation and innovate the way AI agents collaborate to solve practical problems. The CAMEL-AI team provides clear and clear installation tutorials. Whether using Conda or venv, developers can quickly build the OWL operating environment and configure API keys according to the guidelines.
In addition, OWL also provides scripts that reproduce GAIA benchmark test results to facilitate developers to verify their performance. The CAMEL-AI team has announced the future development direction of OWL, including writing technology blogs, enhancing the tool ecosystem, and developing more complex agent interaction models.
By studying Manus' technical route, OWL is actively replicating and surpassing its functions. The CAMEL-AI team has launched the Workflow Replica Plan and the Ubuntu Toolkit Replica Plan. This means that what Manus can do, such as starting Ubuntu containers, connecting to data sources, automatically generating task lists, executing terminal commands, parsing files, operating browsers, etc. OWL is also actively implementing it and will be more flexible.
What is even more exciting is that the technology of CAMEL-AI's previously developed cross-platform agent CRAB will also be integrated into OWL. This means that in the future, OWL will not only be able to control cloud containers, but also hope to directly control applications on mobile phones and computers, real remote operation across multiple devices and all scenarios, and become a veritable open source Manus Pro Max.