Re: [ontolog-forum] RDF finally has its long awaited Generic Client!

Hi Martynas,

On 10/1/25 4:18 PM, Martynas Jusevičius wrote:

> Hi Kingsley,
>
> We can call LLM "a client" but I don't think it can be called "the
> client". Yes they can build static reports and pages and what not, but
> is this as far as we want to push RDF Knowledge Graphs + LLMs? To me,
> that leaves a lot to be desired.

“The Generic RDF Client” — because they add a natural language component 
to computing’s UI/UX stack that can handle the following:

     1.    Generate RDF and/or SPARQL
     2.    Consume RDF and/or SPARQL
     3.    Visualize RDF and/or SPARQL

All of this happens in a loosely coupled manner, rather than being 
hardwired to any backend (DBMS, triplestore, or filesystem).

LLMs are now integrated into browsers, phone apps, and beyond the 
generic chatbot clients (ChatGPT, Claude, Gemini, etc.)—they’re also 
embedded in developer assistant tools like Claude Code, OpenAI Codex, 
and Gemini CLI. They’re everywhere, and wherever they are, they open up 
RDF processing without distraction.

I don’t follow your “static reports” claim, since they can just as 
easily generate highly interactive single-page HTML apps, into which 
content retrieved over HTTP (via various sub-protocols, including 
SPARQL) is dynamically injected.

> I am convinced there needs to be an intermediary management the LLM
> can interact with. Essentially an RDF graph store with a UI.

I disagree.

They simply need to be loosely coupled with Data Spaces (databases, 
knowledge graphs, filesystems, APIs). And that already works, especially 
when you leverage complementary open standards like MCP (Model Context 
Protocol) and OpenAPI.

FWIW — We (OpenLink Software) have built various apps that achieve the 
above. Examples include:

 1.

    OpenLink AI Layer Playground
    <https://www.openlinksw.com/data//html/opal-system-functions-or-agents-apis-developer-journey-gemini.html>
    — for building AI Agents / Assistants that are loosely coupled with
    Data Spaces

 2.

    SPARQLWorks
    <https://www.linkedin.com/pulse/sparqlworks-visualizing-rdf-based-knowledge-graphs-age-heward-mills-oho5e>
    — SPARQL Query and Results Visualization

*“RDF Store with a UI”* is the complete opposite of what I am talking 
about, which clarifies the basis of your pushback. I don’t believe in 
that kind of conflation.

Instead, I prefer a modern *MVC (Model–View–Controller)* pattern, where 
LLMs supercharge the controller by leveraging loose coupling to operate 
on models and generate views [1].

> That unlocks possibilities that your scenario alone doesn’t cover:
> - Persistent storage of generated reports, documents, and RDF data —
> with URIs assigned so they can feed back into the Knowledge Graph
> - Revisiting and updating previously generated content, or querying it directly
> - Dedicated UI for browsing and recall — since chat alone is not the
> most convenient or user-friendly way to manage content. You wouldn't
> want to have the chat as the only interface to Google Drive or Notion
> or something like that?
> - Multi-user collaboration on documents and data

I totally disagree :)

> In short: LLMs can generate RDF artifacts, but without a system to
> manage them, we’re limited to static, one-off outputs. With a proper
> graph store + UI layer, we enable living, collaborative, queryable
> Knowledge Graph applications.

You can build AI agent for which LLMs are simply a component as part of 
the kind of loose coupling that’s now finally possible.

Links:

[1] Software Componentization in the Age of AI: How LLMs Are Reshaping 
Development 
<https://www.linkedin.com/pulse/software-componentization-age-ai-how-llms-reshaping-idehen-xstce/>

Kingsley

> Best,
>
> Martynas
> atomgraph.com
>
> On Tue, Sep 30, 2025 at 5:49 PM Kingsley Idehen<kidehen@openlinksw.com>  wrote:
>> Hi John,
>>
>> On 9/30/25 10:17 AM, John Erickson wrote:
>>
>> I'm confused; what is the actual RDF _client_ we're talking about?
>>
>>
>> Any of the current LLM-based chatbots (ChatGPT, Claude, Gemini, Grok, etc.), or their LLMs via API interactions, can serve as clients that generate and consume RDF simply based on natural language instructions.
>>
>>
>> For example, you can issue prompts like:
>>
>> “Generate a detailed report about {some-event} in RDF using terms from {vocabularies X, Y, Z}.”
>>
>> This works for sporting events like soccer or NFL games, as well as current affairs events.
>>
>> Likewise, you can transform content in non-RDF formats into RDF, or generate SPARQL queries (including SPARQL Protocol URLs). In short, all the bases are covered.
>>
>>
>> Kingsley
>>
>>
>> My "read" of Kingsley's message was more that LLMs facilitate building RDF-consuming clients, and not specifically about _A_ client.
>>
>> If so, where is it? Where's the link to the web app? How do I download it? etc...
>>
>> John
>>
>> On Tue, Sep 30, 2025 at 9:52 AM Filip Kolarik<filip26@gmail.com>  wrote:
>>> On Tue, Sep 30, 2025 at 11:50 AM Alex Shkotin<alex.shkotin@gmail.com>  wrote:
>>>> Hi Kingsley,
>>>>
>>>>
>>>> A good article about using RDF and user interface functionality. But I believe that any information generated by LLM should be marked "May contain errors."
>>>>
>>>> So all those beautiful tables, diagrams, and documents should display this sign prominently.
>>>>
>>>>
>>> I like this client. Thank you, Kingsley, for sharing it. It has its use cases, even though we all know it’s heuristic. That said, it’s still useful for many scenarios, like running quick, conceptual queries to get business data estimates, projections, and insights, etc.
>>>
>>> For other cases, where precise calculations, explainability, and trusted outcomes are required, LLM-based agents could help develop and deliver solutions that are accurate, explainable, and, with proper legal oversight, accountable. The key is combining LLM capabilities with human expertise to balance speed and creativity with accuracy and reliability. I’m not imagining some “AI programmer,” but rather a powerful toolset built on LLMs that effectively leverages knowledge.
>>>
>>> After all, it’s never been about "generating code", but about understanding and clearly communicating the requirements and constraints, that’s the real issue we deal with.
>>>
>>> Best regards,
>>> Filip
>>>
>>>> For me, user interface functionality that reflects the power of RDF is more important.
>>>>
>>>>
>>>> Best regards,
>>>>
>>>>
>>>> Alex
>>>>
>>>>
>>>>
>>>> пн, 29 сент. 2025 г. в 19:48, 'Kingsley Idehen' via ontolog-forum<ontolog-forum@googlegroups.com>:
>>>>> Hi Everyone,
>>>>>
>>>>> It’s been a while!
>>>>>
>>>>> Something important is happening right now, thanks to the emergence of LLMs as the long-awaited generic RDF client (the so-called “killer app”). We all know how Mosaic → Mozilla/Netscape made HTML and HTTP globally usable by end-users and developers alike. Well, the very same thing is finally happening with RDF—albeit some 20+ years later than expected.
>>>>>
>>>>> Here’s a post I recently published on LinkedIn about this critical development:
>>>>>
>>>>> https://www.linkedin.com/pulse/large-language-models-llms-powerful-generic-rdf-clients-idehen-xwhfe
>>>>>
>>>>> --
>>>>> Regards,
>>>>>
>>>>> Kingsley Idehen
>>>>> Founder & CEO
>>>>> OpenLink Software
>>>>> Home Page:http://www.openlinksw.com
>>>>> Community Support:https://community.openlinksw.com
>>>>>
>>>>> Social Media:
>>>>> LinkedIn:http://www.linkedin.com/in/kidehen
>>>>> Twitter :https://twitter.com/kidehen
>>>>>
>>>>>
>>>>> --
>>>>> All contributions to this forum are covered by an open-source license.
>>>>> For information about the wiki, the license, and how to subscribe or
>>>>> unsubscribe to the forum, seehttp://ontologforum.org/info
>>>>> ---
>>>>> You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
>>>>> To unsubscribe from this group and stop receiving emails from it, send an email toontolog-forum+unsubscribe@googlegroups.com.
>>>>> To view this discussion visithttps://groups.google.com/d/msgid/ontolog-forum/9501caa2-29b1-4092-8866-db47c0c23cc1%40openlinksw.com.
>> --
>> Regards,
>>
>> Kingsley Idehen
>> Founder & CEO
>> OpenLink Software
>> Home Page:http://www.openlinksw.com
>> Community Support:https://community.openlinksw.com
>>
>> Social Media:
>> LinkedIn:http://www.linkedin.com/in/kidehen
>> Twitter :https://twitter.com/kidehen
>>
>>
​

-- 
Regards,

Kingsley Idehen 
Founder & CEO
OpenLink Software
Home Page:http://www.openlinksw.com
Community Support:https://community.openlinksw.com

Social Media:
LinkedIn:http://www.linkedin.com/in/kidehen
Twitter :https://twitter.com/kidehen