New AI-powered features on the THETA Developers Forum

New AI-powered features on the THETA Developers Forum

We are excited to introduce new AI-powered features on the THETA Developers Forum as part of a limited experimental phase. During this period, we are actively testing and refining these capabilities, which means some features may change or be temporarily disabled based on feedback and performance. Additionally, certain usage quotas may not yet be properly configured, so we appreciate your patience as we fine-tune the experience.

Several of these new features, such as AI Chatbot, AI Helper and Topic Summarization, require users to log in to access them. We encourage you to create a free account, sign in, and try them out.

AI Chatbot

The AI Chatbot pulls from information in the forum but provides a way to ask questions privately, when needed.

AI Helper

The AI Helper allows you to translate your text to English, proofread your post, generate a table in Markdown, and add smart dates.

AI Summarize

Summarize gives you the ability to quickly make a summary of the key information from a full thread, often covering multiple responses and different opinions and covering longer periods of time of accumulated knowledge.

More!

Meanwhile, other AI-driven enhancements—including AI Search, Related Topics, and Topic Summarization—are available without requiring login.

AI Search

Related

Join in! We invite all members to explore these tools and share their feedback to help us improve the forum’s AI integration.

1 Like

One thing to note that is that the chatbot has both a persona and a LLM.

We can improve the LLMs, but there are increase costs. For example, the o1 and o1-mini LLMs from OpenAI appear to be better and are not available in the free version of ChatGPT, the application that most people are familiar with.

Additionally DeepSeek R1 appears to be comparable to o1 at a cheaper price.

When we upgrade the reasoning AIs, it’ll be an even more interesting comparison.

The key difference between an LLM and a Persona:

  • An LLM is the underlying technology
  • A persona is a customization layer that gives the LLM a distinct identity or behavior.

Like a radio station (LLM) that can play different genres of music (personas) depending on what music you like.

So far, in this forum, I’ve used Creative and Helper the most. Honestly, so far, I haven’t noticed a large difference between the two. If anyone tries other Personas and have some feedback, I’d be interested to hear your results/opinions.

Currently, you can choose from these Personas:

Creative
AI Bot with no external integrations specialized in creative tasks

Forum Helper
General purpose AI Bot capable of performing various tasks

DALL-E 3
AI Bot specialized in generating images using DALL-E 3

Discourse Helper
AI Bot specialized in helping with Discourse related tasks

GitHub Helper
AI Bot specialized in assisting with GitHub-related tasks and questions

Researcher
AI Bot with Google access that can research information for you

Settings Explorer
AI Bot specialized in helping explore Discourse site settings

SQL Helper
AI Bot specialized in helping craft SQL queries on this Discourse instance

Web Artifact Creator
AI Bot specialized in creating interactive web artifacts

I’m new to Personas in AI, so I’ve been testing the ones that are available with the AI Chatbot. I asked the same question to all 9 Personas. I’m summarizing and showing the responses here.

High Level: I’m asking a question about using the THETA X API. So I’d like to understand the main steps, and I’d like specific code examples to get me through. I’m using GPT-4 Omni for all of the Personas. I asked this question:

What’s the best way to set the resolution to 11K using the WebAPI on a THETA X?

Conclusions:

  1. Web Artifact Creator, showed the best code examples. First showed JavaScript, responded to prompts for Swift and Kotlin examples
  2. GitHub Helper looked good at a glance but had an important API command wrong. This would be hard to spot for developer just getting used to the API.
  3. Discourse Helper and SQL Helper are intended for questions about the forum software, not about the camera API

Creative

Detailed response, shows curl examples. Adds an Additional Considerations section at the end that is pretty high level and not so useful.

Forum Helper

States that there aren’t any direct results specifically detailing how to set the resolution to 11K using the WebAPI for the THETA X. Then links to 3 threads, 1 of which is directly related. Also links to RICOH official documentation.

DALL-E 3

Shows a general guide to set the resolution to 11K using the WebAPI on a Ricoh THETA X. Says you need to interact with the camera’s API to adjust the capture settings. Shows an example payload for setting camera.setOptions.

Discourse Helper

Searches the Discourse Meta for relevant discussions or guides, which is the site for handling questions around Discourse, the open source forum software that this forum is using. Predictably, it doesn’t find anything.

GitHub Helper

Searches Github for relevant discussions or guides. Says it retrieved content of theta-web-api-v2.1/options/file_format.md from ricohapi/theta-api-specs@main. It produced a concise, step by step way of setting the resolution to 11K. Unfortunately, it somehow came up with an API command that does not exist: photoResolution. When told to correct this, it did, on the first try, showing how to set camera.setOptions.

Researcher

Searches Google, where it found 10 results, and also has “Browse Web” which is listed as one thread on .guide (THETA X API Documentation)

Response had the best summary sentence of all the Personas, in my opinion:

To set the resolution to 11K on a Ricoh THETA X using the WebAPI, you’ll need to utilize the fileFormat option within the API. The 11K image format is supported, and the settings for this can be configured through the WebAPI. But only shows a 'general guideline on how you might set this up."

Settings Explorer

Fairly general response, but gives OK instructions, including showing an example setting camera.setOptions and mentions being able to check the settings using camera.getOptions.

SQL Helper

Just focused on SQL responses and it tells you that: “I’m here to assist with SQL queries and database-related questions, particularly in the context of Discourse forums. For specific technical questions about setting resolutions on devices like the THETA X using a WebAPI, I recommend consulting the device’s official documentation or support resources. If you have any questions related to databases or need help with SQL, feel free to ask!”

Web Artifact Creator

Good response, shows example using JavaScript with the Fetch API. It was easy to ask for Swift and Kotlin examples instead. Key Points section nicely summarizes API Endpoint that you use, API Command that you set, and the Resolution settings. Asks Would you like help setting up a complete web component to interact with the THETA X using this approach? I said Yes and it showed the CSS, HTML and JavaScript to build a button labeled “Set Resolution to 11K” that a user can click to send the API request.