Einrichtungsanfrage: KI-Zusammenfassung in Discourse und LLM-Integration

[S.AAKASH_MUTHIAH]

Hello Discourse Support Team,

I have a few questions regarding the use of AI summarization in Discourse, and I would appreciate your help in guiding me through the setup process:

1. How to Initialize AI Summarization:

   • What steps do I need to follow to activate and configure AI summarization in Discourse?

   • Are there any plugins required to get started with AI summarization? If so, which ones, and where can I find them?

2. Using Our Own LLM (from Dify):

   • Is it possible to integrate our custom LLM, created through the Dify application, for summarization tasks in Discourse? If so, how can we do that using an API?

   • Can you provide any API documentation or integration guides for connecting third-party LLMs to Discourse?

3. Cost of AI Models in Discourse:

• Available LLM Models: What are the different AI summarization models available in Discourse, and how do I choose between them?

• Free vs Paid Models: Are any of the LLM models free to use, or is there a cost associated with all of them?

• Usage Restrictions: Are there any limitations on usage for free models, such as daily request limits or token restrictions?

• Cost Metrics: What is the pricing structure for each model, including any per-request or per-token costs? If there are multiple pricing tiers, can you provide a breakdown for each model?

4. Additional Information:

   • Please let me know if there are any additional considerations or steps I need to take to fully utilize the AI summarization features in Discourse.

Thank you in advance for your help!

[NateDhaliwal]

Hi @S.AAKASH_MUTHIAH , most of your queries can be answered in the docs for the AI plugin here.

---

@S.AAKASH_MUTHIAH

What steps do I need to follow to activate and configure AI summarization in Discourse?

See:

Discourse AI - Summarize

This topic covers the configuration of the Summarize feature of the Discourse AI plugin. Required user level: Administrator Summarize topics and chat channels for a quick recap. Use it in mega topics and large discussions to figure out what’s happening. [Chat channel summarize option] Features Summarize topics from topic map (top and bottom of topic) Summarize Chat channels for specific duration of time (up to 7 days) Cached summaries in prior gen…

And:

Summarize - Gists

This guide explains how to enable and configure Gists, which is part of the Summarize feature of the Discourse AI plugin. Required user level: Administrator This feature isn’t yet available on discourse.org hosting Gists are short AI summaries to help you get a quick overview from topic lists. Features Auto backfills short summaries of topics Toggleable between compact and expanded (with AI summary) topic views Available for all topic views, …

---

@S.AAKASH_MUTHIAH

Are there any plugins required to get started with AI summarization? If so, which ones, and where can I find them?

Yes, the ai plugin. It is included in core if you’re updated, just go to /admin/plugins to enable and configure it.

---

@S.AAKASH_MUTHIAH

Available LLM Models: What are the different AI summarization models available in Discourse, and how do I choose between them?

What LLM to use for Discourse AI?

It’s important to understand the needs of you as the community admin and your members when choosing a Large Language Model (LLM) to power Discourse AI features. Several factors may influence your decisions: Performance for use-case: Are you looking for the best-performing model? Performance can vary depending on the task (e.g., summarization, search, complex reasoning, spam detection). Assessment is based on the model’s ability to generate correct, relevant, and coherent responses. Context …

---

@S.AAKASH_MUTHIAH

Free vs Paid Models: Are any of the LLM models free to use, or is there a cost associated with all of them?

Most are paid, but you can take a look at OpenRouter, which provides some free models as well.

---

@S.AAKASH_MUTHIAH

Usage Restrictions: Are there any limitations on usage for free models, such as daily request limits or token restrictions?

AFAIK this will be up to your provider, but CMIIW you can enforce limits in Discourse as well.

Maybe see:

Estimating costs of using LLMs for Discourse AI

In order to use certain Discourse AI features, users are required to use a Large Language Model (LLM) provider. Please see each AI feature to determine which LLMs are compatible. If cost is a significant worry, one way to combat that is to set usage limits right from the vendor and use a monthly budget. Another option is to only let select users and groups access the AI features There are several variable factors to consider when calculating the costs of using…

---

@S.AAKASH_MUTHIAH

Cost Metrics: What is the pricing structure for each model, including any per-request or per-token costs? If there are multiple pricing tiers, can you provide a breakdown for each model?

That will be up to your AI LLM provider; they differ.

---

See all the docs for the AI plugin here.

For the settings offered in the plugin, see this guide.

---

Hope this helps!

[S.AAKASH_MUTHIAH]

Hi,
Thank you so much for your quick and helpful response!

I checked the plugin options on my Discourse instance, but I don’t see anything related to AI listed there. I’ve attached an image showing the available plugin options on my end — could you please take a look and let me know if I’m missing something?

[NateDhaliwal]

Are you on the latest test-passed version?

[S.AAKASH_MUTHIAH]

my current version

[NateDhaliwal]

Try updating your forum and see if that works. Taking a look at your version, it looks like the AI plugin was not bundled in core yet.

[S.AAKASH_MUTHIAH]

Thank you, It worked now.

I have a question. I created a workflow on DIFY and now I want to use that workflow’s API to integrate my own LLM. Could you also let me know what format the input should be sent in?
And where to add API.
And what option need to be clicked to configure our own.

[S.AAKASH_MUTHIAH]

I found one thing is it this page where i need to configure settings?

If yes what option i should select from this dropdown

[NateDhaliwal]

If you choose ‘Manual configuration’ in the LLMs page, and put your keys in there, does it work?

I haven’t used the AI plugin in a while, so my memory of it is a little hazy.

[S.AAKASH_MUTHIAH]

I would like to understand the type of input that will be sent to the API endpoint from Discourse.