Example Flows

The ZBrain Flow uses a combination of different components to create flow states. These components include:

• Knowledge Bases: Repositories of information that the flow state can access.

• Large Language Models (LLMs): AI models that can be used to generate text, translate languages, and answer questions.

• Sequential Chains Tools that allow the flow state to access information from multiple knowledge sources in sequential order.

• Prompts: Templates that can be used to generate text for the LLMs.

• Agents: Tools that can be used to interact with the real world.

Here are just a few examples of how the ZBrain Flow can be used to create flows.

Default Flow

The flow state is created using these four components: knowledge base, chatOpenAI, LLM chain and memoryBuffer. The knowledge base provides the flow state with the information it needs to generate responses. The chatOpenAI helps the flow state to generate those responses in a natural and engaging way. The LLM chain allows the flow state to create more complex and nuanced responses. And the memoryBuffer allows the flow state to keep track of its progress and state so that it can continue the flow state even if it is interrupted.

Here is an example of how the ZBrain Flow could be used to create a flow state:

• The user asks the flow state a question about a topic that is in the knowledge base.

• The flow state uses the chatOpenAI to generate a response to the question.

The ZBrain Flow can be used to create flow states for a variety of purposes, such as:

• Learning new information

• Solving problems

• Brainstorming ideas

• Generating creative content

Using SequentialChain Flow

The SequentialChain allows the flow state to access information from multiple knowledge sources in sequential order. This information can be anything from facts and figures to stories and poems. ChatOpenAI allows the flow state to generate responses in a natural and captivating manner. Leveraging LLM chains, the flow state can produce responses that are not only more intricate but also finely nuanced. Moreover, the inclusion of knowledge bases empowers the flow state with an extensive array of information to utilize. Combining these knowledge sources allows the flow state to create more comprehensive and informative responses.

Here is an example of how the ZBrain Flow could be used to create a flow state:

• The user asks the flow state a question about a topic that is in two knowledge bases.

• The flow state uses the SequentialChain to access information about the topic from the two knowledge bases in sequential order.

• The flow state uses the chatOpenAI to generate a response to the question.

Request Tool Flow

The requestGetTool provides the flow state with the information it needs to generate responses. The textRequestWrapper helps the flow state to format the information from the requestGetTool in a way that the chatOpenAI can understand. ChatOpenAI helps the flow state create natural and engaging responses. It uses the LLM chain to make more complex and detailed replies. The zeroShotPrompt helps the flow state generate prompts tailored to the specific information it has access. And with the zeroShotAgent, the flow state can safely interact with the real world.

Here is an example of how the ZBrain Flow could be used to create a flow state:

• The user asks the flow state a question about a topic that is in the real world.

• The flow state uses the requestGetTool to access information about the topic from the real world.

• The flow state uses the textRequestWrapper to format the information from the requestGetTool in a way that the chatOpenAI can understand.

• The flow state uses the chatOpenAI to generate a response to the user query.