List of most frequently asked questions.

NeoPilot Frequently Asked Questions

Welcome to the NeoPilot FAQ. Find answers to common questions about creating and managing AI Agents, working with GenAI Pipelines, and using the NeoPilot platform.

Understanding the Basics

What is an Agent?

An Agent is an intelligent software entity that manages user interactions, conversations, and task orchestration. Agents serve as the user-facing component of NeoPilot applications, utilizing widgets for user interfaces and invoking pipelines for complex processing. They maintain conversation context, break down tasks, and present results in user-friendly formats.

What is a GenAI Pipeline?

A GenAI Pipeline is a specialized workflow that processes data using AI models and business logic. Pipelines are the "brain" of the NeoPilot Agentic system, handling complex operations like natural language generation, document analysis, data transformation, and integration with external systems. They operate as reusable, Python-based processing units that can be invoked by multiple agents.

What are the components of Agent Builder?

Components of Agent Builder are the building blocks used to create AI agents. These include:

• Widgets: UI elements like forms, buttons, chat interfaces, and visualizations for user interaction.
  

• Flow Controllers: Logic elements that determine the agent's workflow and decision-making.
  

• Integration Points: Connections to pipelines, external APIs, and data sources.
  

• State Managers: Tools for maintaining conversation context and user session data.

These components can be dragged and dropped in the visual interface Canvas to create sophisticated agent behaviors without extensive coding.

What are API Keys and why are they required?

API Keys are authentication credentials required to access external AI services like OpenAI (GPT models), Anthropic (Claude), Google Maps, and other integrated technologies. They are required because:

• They authenticate your requests to these third-party services.
  

• They enable usage tracking and billing for consumption-based services.
  

• They provide security by ensuring that only authorized applications can access these services.
  z

• They allow rate limiting and quota management for your applications.

NeoPilot uses these external services to power various AI capabilities in your agents and pipelines, which is why API keys are necessary to run most applications.

Getting Started

How do I open the Agent Builder to create an Agent?

To open the Agent Builder:

1.  Log in to your NeoPilot; need to register first.
   

2. Click the App Palette on the top right.
   

3. Select the  WorkSpace from the options.
   

4. In the workspace interface, click the Agents in the left sidebar.
   

5. Click + to create a new agent or select an existing agent to edit.

This opens the visual Agent Builder interface where you can design your agent's workflow and user interface.

For advanced agent-building concepts, developers can reference the Agent Building Component documentation at Agent Builder.

How do I open the Pipeline Builder to build a pipeline?

To open the Pipeline Builder:

1.  Log in to your NeoPilot; need to register first.
   

2. Click App Palette on the top right.
   

3. Select WorkSpace from the options.
   

4. In the workspace interface, click the GenAI Pipelines in the left sidebar.
   

5. Click + to create a new pipeline or select an existing pipeline to edit.

This opens the Pipeline Builder interface where you can design your AI processing workflows.

For advanced agent-building concepts, developers can reference the Agent Building Component documentation at GenAI Pipeline Builder.

How do I trigger a pipeline inside an Agent?

To trigger a pipeline from within an Agent:

1. Open your agent in the Agent Builder.
   

2. Add a GenAI Pipeline Builder agent-building component to your agent's workflow.
   

3. In the component's properties panel, select the pipeline you want to trigger.
   

4. Configure input parameters that are sent to the pipeline (these can be static values or dynamic values from user input or agent state).
   

5. Specify how to handle the pipeline's response (such as updating the agent state, displaying results, make decisions).
   

6. Connect this component to appropriate triggers.

When the agent reaches this point in its workflow, it automatically invokes the pipeline and waits for results before continuing.

How many pipelines are needed to create one Agent?

There is no fixed requirement for how many pipelines are needed to create an agent. It depends entirely on your application's complexity and functionality:

• Simple agents might need just one pipeline or even none at all if they primarily handle UI interactions.
  

• Moderate agents typically use 2-5 pipelines for different functions (e.g., data analysis, content generation, image processing).
  

• Complex agents can invoke 5+ pipelines for sophisticated workflows with multiple specialized tasks.

The modular design of NeoPilot allows you to start with a minimal or zero set of pipelines and add more as your agent's capabilities expand. You can also reuse the same pipeline across multiple agents if they share similar processing needs.

Running and Managing Agents

How do I deploy/publish an Agent?

To deploy/publish an Agent for production use:

1. Once your agent is ready and fully configured. Click PUBLISH (previously labeled as Unpublish) located at the top section of your agentic workflow. Then, click SAVE to finalize your changes.
   

2. Navigate to NeoPilot Agentic AI Engine from the application palette.
   

3. Click Agents. Under the My Agents section, you can find a list of all your published agents.
   

4. Select the desired agent from the list to launch and run it in the production environment.

Where can I view my ready Agents?

You can view all your ready Agents in several places:

• WorkSpace Library: Shows all agents you have created or have access to.

• Deployment Console: Lists agents that have been deployed for production use.

These interfaces provide status information, usage statistics, and management options for your agents.

How do I run my Agent?

To run your Agent:

1. Navigate to NeoPilot Agentic AI Engine from the application palette.
   

2. Click Agents. Under the My Agents section, you can find a list of all your published agents.
   

3. Select the desired agent from the list to launch and run it in the production environment.

To clarify in detail, the user can refer to topic 4. Deployment and Maintenance. documentation at Getting Started with Agent Builder.

I can not see my GenAI pipeline in Agent Builder. What should I do?

If you can not see your GenAI pipeline in the Agent Builder, try these troubleshooting steps:

1. Check the access level of your pipeline - it must be set to at least Shared for you to access it in Agent Builder.
   

2. Verify that the pipeline has a description and that it has been saved successfully - unsaved or draft pipelines do not appear.
   

3. Refresh the Agent Builder interface - sometimes the pipeline list needs to be refreshed.
   

4. Ensure you are logged in with the same account that has access to the pipeline.
   

5. Check if the pipeline is in the same project/organization context as your agent.

To learn more about building pipelines, check out the GenAI Pipeline Builder. For guidance on integrating a GenAI pipeline into an agentic workflow, refer to the Call Task Pipeline component documentation under Agent Building Components.

If these steps do not resolve the issue, try logging out and back in, or contact your system administrator for assistance with pipeline permissions.

API Keys & Tokens

How do I purchase an OpenAI key?

To purchase an OpenAI API key:

1. Visit the OpenAI website at https://openai.com/api/
   

2. Create an account or log in to your existing account.
   

3. Navigate to the API section.
   

4. Select a pricing plan that suits your needs.
   

5. Provide payment information and complete the purchase.
   

6. Once payment is processed, you receive your API key.
   

7. Copy this key and add it to your NeoPilot account settings in the "API Keys" section.

OpenAI typically offers a pay-as-you-go model with different rates for different models. They may also provide some free credits for new users to test the service.

How do I purchase a Claude key?

To purchase an Anthropic Claude API key:

1. Visit Anthropic's website at https://www.anthropic.com/product
   

2. Follow the instructions to request API access (Claude API access may require approval).
   

3. Once approved, create an account or log in.
   

4. Navigate to the API section in your account.
   

5. Select a pricing plan appropriate for your usage.
   

6. Complete the payment process.
   

7. Retrieve your API key from your account dashboard.
   

8. Add this key to your NeoPilot settings in the API Keys section.

Note: Anthropic's API access policies and pricing may change, so check their current requirements on their website.

How do I purchase a Google Maps key?

To purchase a Google Maps API key:

1. Go to the Google Cloud Platform Console at https://console.cloud.google.com/
   

2. Create a new project or select an existing project.
   

3. Navigate to APIs & Services > Library.
   

4. Search for and enable the Google Maps APIs you need (Maps JavaScript API, Places API)
   

5. Go to APIs & Services > Credentials.
   

6. Click Create credentials and select the API key.
   

7. Set up billing for your Google Cloud project.
   

8. Configure usage restrictions and quotas as needed.
   

9. Copy your API key and add it to your NeoPilot account in the API Keys section.

Google typically offers a free tier with some usage included before charging, Review their current pricing model for the specific APIs you plan to use.

How do I purchase a Gemini key?

To purchase a Google Gemini API key:

1. Visit the Google AI Studio website at https://ai.google.dev/gemini-api/docs/api-key
   

2. Sign in with your Google account.
   

3. Click Get API Key or navigate to the API access section.
   

4. Agree to any terms of service and select your usage plan (some tiers may offer free access or require a billing setup).
   

5. Once generated, copy your Gemini API key.
   

6. Add this key to your NeoPilot account in the API Keys section.

Google Gemini may require you to enable billing or link your account to a Google Cloud project for production-level use. Be sure to review Google’s documentation for usage limits, pricing tiers, and API capabilities.

Why do I need to purchase an API key to run an Agent?

API keys need to be purchased for several important reasons:

• Third-party services: NeoPilot integrates with external AI services like OpenAI, Claude, and Google Maps. These companies charge for their computing resources and AI capabilities.
  

• Computing costs: AI models require significant computational resources to run, and the companies that develop them charge to cover these costs.
  

• Fair usage model: The pay-per-use model ensures you only pay for what you use, making advanced AI accessible without massive upfront investments.
  

• Ongoing improvement: Revenue from API usage helps these companies continue improving their AI models.

While NeoPilot provides the platform and tools for building AI agents, the actual AI processing often occurs on these third-party services, which is why their API keys are required.

How many tokens are required to build a business application?

The number of tokens required to build a business application varies widely depending on several factors:

• Application complexity: Simple Q&A agents might use 10,000-50,000 tokens per day, while complex data analysis applications could use millions.
  

• User volume: More users means more interactions and higher token usage.
  

• Interaction depth: In-depth conversations require more tokens than simple queries.
  

• Content generation: Generating long-form content (reports, articles) uses more tokens than short responses.

As a rough estimate:

• Small business application: 100,000 - 500,000 tokens per month.
  

• Medium business application: 500,000 - 2,000,000 tokens per month.
  

• Large enterprise application: 2,000,000+ tokens per month.

Most providers charge per 1,000 tokens, with rates varying by model. It's advisable to start with a smaller allocation and scale up as you monitor actual usage patterns.

Marketplace

I want to sell my Agent. How do I upload it to the Marketplace?

To sell your Agent on the NeoPilot Marketplace:

1. Prepare your Agent: Ensure it is fully functional, well-documented, and thoroughly tested.
   

2. Access the Marketplace Provider Portal: From your dashboard, navigate to Marketplace > Become a Provider.
   

3. Create a listing: Fill out the listing form with:

   • Title and description of your Agent.

   • Screenshots and/or demo videos.

   • Pricing model (one-time, subscription, usage-based).

   • Category and tags for discoverability.

   • Documentation and support information.
     

4. Set up payment details: Configure your payment information to receive sales revenue.
   

5. Submit for review: Marketplace submissions undergo a review process to ensure quality and security.
   

6. Address feedback: Make any necessary changes based on the review team's feedback.
   

7. Publish: Once approved, your Agent is published to the Marketplace.
   

8. Manage your listing: Monitor sales, respond to reviews, and update your Agent as needed.

The review process typically takes 3-5 business days. Higher-quality submissions with complete documentation and impressive capabilities are more likely to be approved quickly.