Welcome

What is Fungi Studio?

Fungi Studio is Agent Development Environment (ADE) used to build advanced AI Agents with DeFi Capabilities. With our generalized user-interface, the studio simplifies how we interact with and harness Large Language Models (LLMs). It provides a visual, graph-based interface for creating, connecting, and orchestrating AI agents that can perform complex tasks through carefully crafted prompts and carefully routed workflows.

What is an Agent Development Environment?

Most existing frameworks do no live up to the promises of an ADE, therefore, an ADE is a framework but with more prerequisites:

• Manageable file system: in the studio, you get to create how every many separate files to accommodate your specific use case.

• Custom dependencies: for each project, if you require custom dependencies, we will spin up your own private environment to execute custom scrips in either Python or JavaScript.

• Custom scripts: You can provide agents built in our studio with any custom executable code, with access to any supported packages or APIs.

• Rearrangeable workflow designs: with powerful deterministic or inference based routing, your agentic workflows can be as vast as you design, with even the option to nest reusable graphs within single nodes.

• Iterative through branch/version control: We have tailored a robust set of versioning features inspired by Git. Control how many branches of your project exist, while being able to test and execute different versions of the same graph before merging to your main branch so that your project can stay protected from large changes.

What kind of agents can I develop on Fungi Studio?

There are two generalized use cases for an agent hosted through the studio. They will be either autonomous or triggered by client interactions.

Autonomous

These agents are designed to run on the background executing jobs on behalf of a user.

There are several different types of executions you can imagine an autonomous agent doing:

• Completing a DeFi transaction

• Performing automated long-term research

• Sending out notifications for specific data responses

See the Mushy docs as an example for an autonomous DeFi agent. And try the live Mushy app to see a fully realized application deployed entirely on the studio.

Client Interfacing

These agents are designed to be interactive either via a simple chat interface or other client interactions such as button clicks, form completions, etc.

There are several different types of use cases for developing client-based agents:

• Chat with an assistant AI for a product

• Trigger a summarizing agent after the completion of a form

• Design an AI to automatically reply to social media posts

How can I get started?

Fungi Studio is currently available via a closed beta and you can apply for this exclusive access by completing our beta request form.

Certain studio features are locked behind specific access requirements that can be opened by the Fungi team upon request, so if you have any issues or concerns with your current studio project, please reach out to [e-mail]

The agents are waiting.