Apologist Agent is a sophisticated AI chatbot that has been developed to provide accurate, nuanced, and empathetic responses to complex theological inquiries. The chatbot serves two primary audiences:
- Individuals who do not identify as Christians but are earnestly seeking answers to profound theological questions. While it is always preferable for seekers to consult with trained clergy members, it must be acknowledged that many individuals may hesitate to engage in such conversations due to fear of potential confrontation or because of limited access to qualified people to engage with.
- Christians who are interested in deepening their theological knowledge or who are searching for more effective strategies to engage in meaningful dialogue with non-Christians they wish to minister to. Apologist Agent can serve as a valuable resource for these individuals, providing insights and guidance to help them navigate challenging conversations about faith.
Prompt Input #
Apologist Agent (or simply Agent) currently supports three prompt input methods:
- Written Prompt
- Spoken Prompt
- Randomly Selected, Predefined Prompt
Written Prompt #
The primary way for you to interact with Agent is by typing out a prompt and clicking on the purple sparkle button just to the right of the text input.
![](https://apologistproject.org/wp-content/uploads/2024/05/Written-Prompt-1024x156.png)
Spoken Prompt #
Agent also supports voice input. Simply click on the microphone button. The microphone button will pulse while it listens to what you say. Your browser may prompt you to allow access to your computer’s microphone. Once Agent detects that you have finished what you are saying, it will enter what it heard into the text input. From there, you may correct the text before submitting it by clicking on the purple sparkle button.
![](https://apologistproject.org/wp-content/uploads/2024/05/Random-Pre-Defined-Prompt-300x170.png)
Randomly Selected Predefined Prompt #
If you’re new to Agent, testing it out, or simply unsure of what to ask, simply click on the down arrow when Agent first loads. This will randomly select a predefined prompt and automatically fill the prompt input field for you, making it easy to explore Agent’s capabilities and discover new ways to interact with the chatbot.
Response Output #
![](https://apologistproject.org/wp-content/uploads/2024/05/Screenshot-2024-05-03-at-10.16.58 AM-300x298.png)
Apologist Agent is trained to provide biblically-based responses that incorporate Bible verses and context from trusted extra-biblical sources. To get a sense of the range of extra-biblical materials Agent utilizes, you can refer to the Context Source List. This list is continuously expanding as publishers, apologetics ministries, and individual apologists submit their content for consideration through Apologist Ignite.
Whenever Agent’s response includes information from extra-biblical sources, those sources are properly cited and can be accessed by clicking the “Further reading on this topic” button at the end of the response. This button provides users with a deeper dive into the cited sources, offering more details and links to explore the original materials. As a result, any content uploaded through Apologist Ignite that is cited in an Agent response benefits from increased visibility and exposure.
Prompt & Response Lifecycle #
Apologist Agent uses a technique called Retrieval Augmented Generation (RAG) to pass relevant context to a Large Language Model (LLM), which in turn derives its response based on that context. The full prompt lifecycle is outlined below:
- User enters a prompt into the Conversational Interface (Agent).
- Conversational Interface calls the Conversational API.
- Conversational API queries the Corpus of Licensed Materials (Extra-Biblical Sources).
- Corpus of License Materials responds back to the Conversational API with snippets of context.
- Conversational API packages the additional snippets of context along with the original prompt and sends it all to the Large Language Model (LLM).
- Large Language Model responds to prompt given the passed context to the Conversational API.
- Conversational API packages the response and references to the Corpus of Licensed Materials and send them back to the Conversational Interface.
- Conversational Interface displays the response and references to the Corpus of Licensed Materials to the User.
Conversations #
![](https://apologistproject.org/wp-content/uploads/2024/05/Spoken-Prompt-1-300x174.png)
Apologist Agent understands the context of followup questions. Each time a new prompt is submitted, Agent will incorporate the context from the last exchange (your previous prompt and its answer). For example, you can followup with “Explain that again in a different way.” — and Agent will know what you’re referring to.
However, at times you may wish to change the subject and start a new conversation entirely. You may do so by clicking on the button to the far right of the prompt panel at the bottom of the screen. Doing so will wipe out the exchanges, and your conversation can begin again with a clean slate.
Switching the Language #
![](https://apologistproject.org/wp-content/uploads/2024/05/Screenshot-2024-05-03-at-10.59.19 AM-300x132.png)
Apologist Agent officially supports ten languages: English (default), Spanish, Chinese, French, German, Italian, Portuguese, Dutch, Japanese, and Russian. If you would like to change the language of Agent’s responses, click on the flag icon in the top right of the screen. A list of languages to choose from will be displayed. Once you click on a new language, Agent will be instructed to respond using that language.
Please note, however, that non-English responses are not as predictable as English ones. The underlying Large Language Model may at times weave in and out of English and the chosen language.