This week’s discussions on meta.discourse.org revolved around exploring new AI models, configuring AI-powered spam detection, managing AI personas and prompts, and leveraging semantic search capabilities. Users shared their experiences with various language models, discussed potential improvements, and sought guidance on optimizing AI features within the Discourse platform.
Interesting Topics
Don initiated a discussion on Trying out the new AI models, sharing their experiences with models like grok-2-1212 and Gemini Flash 2. Challenges with language detection, formatting issues, and search capabilities were highlighted.
Sam announced a new AI-powered Spam detection module in Discourse AI, providing better control over spam scanning. They shared a configuration guide and encouraged users to share their experiences.
MarcP sought assistance with the Semantic Search API, and Sam provided guidance on accessing pure embeddings search and using API keys to bypass rate limits.
Don requested a Topic AI Summary box closing feature, allowing users to close the summary box by clicking outside it, which awesomerobot acknowledged as a reasonable request.
roniw raised concerns about privacy and data access when using Discourse’s default LLM providers, to which Sam responded, assuring that Discourse LLMs are self-hosted and data is not shared with third parties.
Don initiated the discussion on Trying out the new AI models, sharing their experiences and seeking advice from MihirR and Sam on improving language detection and search capabilities.
roniw raised concerns about privacy and data access when using Discourse’s default LLM providers, and Sam provided assurances about Discourse’s self-hosting and data privacy practices.
This week on the Discourse meta forum, discussions centered around the AI-powered spam detection feature, the proposed “Solved Finder” tool, issues with AI summarization, and various bug reports and feedback related to AI capabilities. Users shared insights, reported bugs, and suggested improvements, fostering a collaborative environment for enhancing the AI integration within the Discourse platform.
Interesting Topics
AI powered Spam detection: Users discussed the rollout of the AI-powered spam detection feature, including confusion around enabling it on hosted accounts and issues with the toggle behavior. @sam provided clarification on the process, and @Saif mentioned an upcoming fix for the toggle issue.
Solved finder: @Don proposed adding a button for staff users to find potential solutions within a topic, similar to the AI summary feature. @sam suggested exploring custom prompts or a “Talk to my topic” option to address this need.
AI Summary Backfill is wasting tokens summarizing PMs: @markschmucker raised a concern about the AI Summary Backfill process summarizing private messages (PMs), which could be inefficient. @Falco acknowledged the issue and mentioned the possibility of introducing a setting to skip PMs, similar to the existing setting for embeddings.
Can’t change ai_embeddings_model: @Overgrow reported an issue with changing the ai_embeddings_model setting and enabling embeddings via the admin panel. @sam provided a link to a pull request that aimed to fix the problem.
How Does Regenerating Summaries Work?: @markschmucker shared their experience with regenerating summaries using a different LLM and the challenges faced when the backfill process stopped after reaching the specified age limit, despite some topics failing due to rate limits.
Feedback on AI Summarisation: @markschmucker sought clarification on addressing invalid summaries, referring to a previous comment by @sam about potential fixes through refactoring or adjusting personas.
@MachineScholar shared their perspective on the dilemma of summarizing PMs, suggesting that the decision should be based on the community type and the potential value of generating embeddings for PMs.
@markschmucker shared their experience with regenerating summaries using a different LLM and the challenges faced when the backfill process stopped after reaching the specified age limit, despite some topics failing due to rate limits.
The past week on meta.discourse.org saw a diverse range of discussions around Discourse’s AI features, with users sharing their experiences, providing feedback, and seeking support. The community explored the integration of Discord and Discourse, the potential of AI-powered Q&A bots, and the capabilities of AI-powered spam detection. Additionally, there were conversations around the new Smart Dates feature in Helper, the AI Persona Editor, and the availability of AI features on Standard and Business plans.
Interesting Topics
Do you use Discord and Discourse? sparked a discussion about the simultaneous use of Discord and Discourse, their respective strengths, and the potential for an AI-powered Q&A Discord bot that leverages Discourse as a knowledge base.
Write out smarter dates with AI introduced the new Smart Dates feature in Helper, which converts human-typed times and dates into Discourse-compatible, timezone-friendly formats.
AI powered Spam detection highlighted the efficiency of Discourse’s AI-powered spam detection system, with users reporting positive experiences in identifying and hiding spam posts.
Discourse AI - AI triage explored the AI triage feature, which automates the categorization of topics using AI models, leading to discussions around the scalability and configuration of this feature.
Proofreading is too creative highlighted an instance where the AI proofreading feature in Helper made overly creative suggestions, prompting a discussion on the appropriate level of creativity for such features.
Feature requests for AI-Summarization gathered requests for improvements to the AI Summarization feature, including prioritizing the summarization of recently updated topics and the ability to skip summarizing personal messages.
How to hide bot completely, for everyone addressed a user’s request to hide the Discourse AI bot from their community, leading to a discussion on the limitations and potential solutions.
Discourse AI - Helper continued the ongoing discussion around the Helper feature, with users providing feedback and seeking clarification on its functionality.
This week on the meta.discourse.org forum, discussions revolved around various aspects of the Discourse AI plugin, including new features, improvements, and user queries. The AI-powered spam detection tool garnered attention, with users seeking clarification on its configuration and usage. Additionally, the “Ask Discourse” feature, which allows users to interact with an AI persona for documentation assistance, was introduced. Other topics included sortable sentiment tables, handling detailed content with AI, regenerating topic summaries, and managing AI usage quotas.
Interesting Topics
@per1234 highlighted an issue where the default LLMs were not selectable on the ‘Spam’ tab, despite being configured. @sam acknowledged the problem and provided a solution.
@Saif introduced the Ask Discourse feature, which allows users to interact with an AI persona to search through documentation and answer common questions related to Discourse.
@noahl requested the ability to sort sentiment tables by highest to lowest numbers, making it easier to identify and address common sentiments.
@c12gene raised a concern about the AI bot’s inability to read and process detailed content encapsulated in [summary] tags, prompting discussions on prompt engineering and tool usage.
@awesomerobot provided an update on the issue of the AI summary box closing, with a link to the relevant pull request.
@EricGT shared insights on LLM prompt evaluations, emphasizing the importance of understanding prompt effectiveness for LLM development.
@sam announced an upcoming feature for hard caps on user group AI usage for AI bots and AI Helper, allowing admins to define limits on total tokens and requests per group.
@per1234 reported a bug where posts and accounts were not always restored when a flag from the Discourse AI spam detection was rejected.
@Moin sought guidance on hiding the bot completely from all users, prompting a discussion on the visibility of the bot in search results.
Activity
@Saif introduced the Ask Discourse feature and provided details on its functionality and accessibility in the initial post. @Ed_Sshared their experience with the resource and reported a broken link, which @Saif later updated.
@EricGTshared insights on the importance of LLM prompt evaluations for effective prompt engineering.
@per1234reported a bug where posts and accounts were not always restored when a flag from the Discourse AI spam detection was rejected.
@Moinsought guidance on hiding the bot completely from all users, prompting a discussion with @Saif on the visibility of the bot in search results.
@BrianCinquired about the possibility of allowing users to upload documents for the AI bot to read and process, and @Saifclarified that this use case is currently not supported.
@MarcPasked about updating summaries for single post topics that are frequently edited.
Thanks for reading, and I’ll see you again next week!
This week on the meta.discourse.org forum, discussions revolved around various AI-related topics, including a proposed moderation tool for formatting code using AI, the ongoing development of Discourse AI features, issues with AI summarization backfills, and the potential for uploading and discussing PDFs within the composer. Additionally, there were conversations about setting usage limits for AI, evaluating costs between different AI providers, handling AI-generated spam, and exploring the capabilities of AI bots and custom tools.
Interesting Topics
@merefield raised an idea for a moderation tool that would allow trusted users to format code blocks using AI, potentially improving readability and assisting new users who struggle with proper code formatting.
Discussions continued on the Discourse AI plugin, with @sam expressing interest in allowing users to upload large files and ask questions about the content using a persona-based approach.
@markschmucker encountered an issue where the AI summarization backfill process kept regenerating summaries for the same topic, even after a valid summary was already present. This led to a fix by @Roman_Rizzi to make the job more resilient.
@BrianC proposed a feature request to allow users to upload PDFs or text files directly in the composer and have the AI process and respond to questions about the content.
There was a discussion around setting per-group token and usage limits for AI features, with @sam clarifying that quotas are defined per group and applied per user, rather than being shared among users.
There was a bug report regarding posts and accounts not being restored when flags from Discourse AI spam detection were rejected, which @sam addressed with a fix.
Discussions took place around self-hosting embeddings for DiscourseAI, with @sam mentioning ongoing work to restructure the embedding configuration and plans to support multi-model embeddings.
@David_Ghost inquired about the ability of AI Triage to perform searches and avoid topics with similar titles based on creation dates.
@David_Ghost inquired about the ability of AI Triage to perform searches and avoid topics with similar titles based on creation dates, with @sam indicating that such “agent-like” behaviors are being considered.
@c12gene reported an issue with the AI bot not being able to read summaries and detailed content, which was addressed by @MachineScholar through system prompt improvements.
@BrianC asked about tying token limits to subscriptions and allowing more expensive models to be used for a fee, with @sam confirming that different quotas can be set for different user groups.
This week on meta.discourse.org, the Discourse team continued to enhance and refine the AI capabilities within the platform. Key discussions revolved around improving the AI summarization feature, managing costs through usage quotas, and integrating new language models. The community also explored potential use cases, such as formatting code with AI assistance and enabling document uploads for AI analysis.
Interesting Topics
Translate Discourse automatically (without a button) (ref)
The team provided updates on their progress towards enabling automatic translation of Discourse topics using AI models. While the initial target is to have the topic page translatable via a language toggle, challenges around search functionality and handling multilingual content remain.
A new feature was released that allows administrators to define usage quotas for AI models, enabling better cost control and fair access to AI features across the community.
sam proposed the idea of allowing certain user groups access to an “AI Helper” feature for formatting source code in posts, acknowledging the potential performance trade-offs.
AI Plugin Causes All Posts to Be Unreadable in Latest Discourse Version (ref)
shannon1024 reported a critical bug where the AI plugin rendered all posts unreadable after updating to the latest Discourse version. The issue was related to the AI embeddings configuration and was resolved by disabling the relevant setting.
Saif provided guidance on factors to consider when choosing a Large Language Model (LLM) for Discourse AI, including performance, context length, compatibility, language support, multimodal capabilities, and speed.
DeepSeek provider support? What to do when model provider isn’t in “Provider” list? (ref)
MachineScholar inquired about integrating the DeepSeek R1 model, which Falco provided a solution for by configuring it as an OpenAI model.
Saif introduced the AI usage page, designed to help administrators understand how the community is utilizing Discourse AI features over time, aiding in cost estimation and management.
shannon1024 reported a critical bug where the AI plugin rendered all posts unreadable after updating Discourse, which was resolved by disabling the AI embeddings setting.
Saif provided guidance on factors to consider when choosing an LLM for Discourse AI, such as performance, context length, compatibility, language support, multimodal capabilities, and speed.
This week on meta.discourse.org, discussions centered around AI-related issues, features, and integrations. Key topics included debugging AI chat, configuring AI spam detection, utilizing reasoning models like DeepSeek-R1, and exploring the capabilities of the AI Helper plugin. Users shared insights on cost comparisons, performance evaluations, and potential use cases for various language models.
Interesting Topics
dsims encountered an issue with the AI Helper’s post illustration feature, leading to a discussion on updating the required settings. [ref](AI-Helper post illustration error. ai_openai_api_key missing)
oppman sought guidance on integrating AI with external sites and GitHub repositories for a software developer community. [ref](AI Integration with Specific External Sites and GitHub)
Users discussed the cost and performance of the DeepSeek-R1 reasoning model compared to other language models like GPT-4 and Claude. [ref](DeepSeek provider support? What to do when model provider isn’t in “Provider” list?)
MachineScholar reported a random error when interacting with the DeepSeek-R1 model via the AI Bot. [ref](DeepSeek-R1 randomly producing “Job exception: undefined method `finish’ for nil” error)
Eric_Platzek sought assistance with debugging AI chat and clarifying the usage and costs associated with different language models. [ref](Debugging AI chat)
markschmucker shared their experience with the AI Web Artifacts feature and inquired about editing generated code and starting new contexts with existing artifacts. [ref](Announcing: AI Web Artifacts)
NKERIFAC_CLAUD_NBAPNON encountered an issue with the Discourse AI embeddings configuration after an update, leading to a discussion on troubleshooting and configuration steps. [ref](Problems with Discouse AI embeddings configuration)
Activity
sam provided insights on exporting forum posts for manual upload into external language models using the Data Explorer. [ref](Exporting all Forum Posts for Manual Upload into External LLMs?)
Roman_Rizzi shared updates on resolving issues related to the AI plugin causing posts to be unreadable in the latest Discourse version. [ref](AI Plugin Causes All Posts to Be Unreadable in Latest Discourse Version)
BrianC inquired about managing uploaded text files in AI personas and the possibility of editing indexed content. [ref](Discourse AI Persona, upload support)
per1234 reported an issue where posts and accounts were not always restored when flags from the Discourse AI spam detection were rejected, leading to a discussion on potential solutions. [ref](Posts and account not always restored when flag from Discourse AI spam detection rejected)
dsims suggested adding a quality parameter for the Dall-E 3 image generation feature to reduce costs. [ref](Image sizes of Dall-E 3?)
Over the past week (2025-02-03 to 2025-02-10), AI discussions on meta.discourse.org have been both vibrant and diverse. Active debates ranged from fine-tuning the AI helper’s ability to distinguish between Discourse and Discord to exploring innovative ways to pass data to artifacts and chain multiple AI triage scripts. Users focused on enhancing features (as seen in discussions like AI helper does not know the difference between Discourse and Discord and post by sam in Support and ai) while also addressing bugs and configuration issues in sentiment analysis and semantic search. In-depth posts covering topics such as flexible artifact embedding (How do I pass data to an artifact? through post 7) and automated AI workflows (AI + Automation Governance: Orchestrating Independent AI Triage Scripts to post 8) demonstrated the community’s commitment to practical innovation. Meanwhile, bug reports such as issues with sentiment displays and semantic search anomalies were met with proactive troubleshooting by experienced users. These discussions, backed by numerous posts and detailed responses, highlight a week of forward-thinking ideas balanced by diligent debugging.
Interesting Topics
AI Helper Confusion (ai, Support): RGJ raised concerns in post 1 about the AI helper mixing up Discourse with Discord, while sam elaborated in post 2 on potential prompt enhancements.
AI Summarization Backfill Update (ai, ai-summarize): Roman_Rizzi confirmed improvements in the summarization process via post 18, highlighting the use of last_posted_at and revision checks.
Discourse AI Plugin Settings Query (official, ai, Plugin):
In post 193 of the “Discourse AI” topic, Bathinda queried the missing setting for topic mentions, prompting further discussion.
Semantic Search Anomaly (ai, ai-search, Support): tyler.lamparter reported challenges with AI semantic search in post 1, noting discrepancies between AI-powered and normal search results.
DeepSeek-R1 Error Resolution (ai, Bug): MachineScholar confirmed in post 4 that DeepSeek-R1 now functions reliably despite intermittent 502/504 errors from external endpoints.
This week on meta.discourse.org the AI community engaged in in‐depth technical discussions and experiments with several new and evolving features. Contributors debated the feasibility of extending RAG to support diverse PDF formats—highlighting nuances in text extraction and OCR reliability (ref, ref)—and launched an intriguing one‐click experiment for AI summarization with mixed impressions (ref, ref). There were spirited conversations around ChatGPT’s role in forum assistance (ref, ref), together with detailed troubleshooting sessions for semantic search issues that consistently returned irrelevant results (ref, ref). Additional topics—including problems with rake tasks in the AI plugin (ref, ref), automation governance experiments (ref, ref), and fixes for persona tool JSON errors (ref, ref)—kept the forum buzzing with ideas and actionable fixes. Throughout these discussions, top contributors like sam, Yenwod, Jagster, Saif and many others drove the conversation forward by sharing clear troubleshooting steps, potential feature improvements, and critical reflections on AI integration in forum functionalities.
Interesting Topics
Will RAG Support PDF Files in the Future?
Users debated the challenges of supporting all types of PDFs in RAG workflows. MachineScholar sparked the discussion with a comment praising the commit (ref), while Saif explained the intricacies of handling text versus image PDFs (ref). Subsequent troubleshooting steps were detailed by Yenwod when he encountered indexing delays (ref) and further debugging was shared (ref, ref, ref).
New Experiment: Enable AI Summarize on your Discourse with one-click!
In this experiment, users evaluated a one-click activation for summarizations. shooj raised an initial query about limiting the feature to admins or mods (ref), and Jagster argued over its universal benefits (ref). Contributions by Arkshine provided CSS-based workarounds (ref), while additional observations by Jagster and Carverofchocie sparked extra debate (ref, ref, ref, ref).
How are we all feeling about ChatGPT and other LLMs and how they’ll impact forums?
The conversation here balanced optimism with caution. Tris20 offered insights on repetitive answers by LLMs in forum settings (ref), while Saif looked at future directions for LLM assistance in topic formulation (ref).
Getting a lot of no results for semantic search
A technical debate emerged when users reported that AI search always returned roughly 40 results regardless of query relevance. sam began the discussion (ref), and tyler.lamparter shared his observations (ref, ref). Further elaboration on model behaviors and result uniformity was provided by sam and Falco (ref, ref, ref, ref).
Rake tasks in the AI plugin not working
A practical issue was raised by Yenwod regarding failures in AI rake tasks, especially with ruby-progressbar related errors (ref). His follow-up post provided more context (ref), while sam and Yenwod further discussed potential remedies (ref, ref).
AI + Automation Governance: Orchestrating Independent AI Triage Scripts
This thread explored automating workflows for AI triage. Cloud_spanner shared a detailed workflow proposal (ref), with sam suggesting a custom tool for triage automation (ref), and Cloud_spanner reinforcing the benefits of a flexible IFTTT-style approach (ref).
AI Persona using Categories tool – “An empty string is not a valid JSON string”
An error in using the Categories tool left users troubleshooting persona failures. markschmucker consistently encountered the JSON error (ref), while sam confirmed the bug and soon provided a fix (ref, ref).
All AI functions are working ok, but AI search gives 500 error
A disruptive 500 error in AI search caught attention on this thread. Bathinda reported the initial error (ref) and subsequent status updates (ref, ref). Jagster guided users through troubleshooting with actionable solutions (ref, ref, ref).
Discourse AI – Spam detection
Focusing on content quality, this topic examined AI’s capacity for spam detection. jordan-violet shared initial testing scenarios (ref, ref), while Jagster critiqued the logical phrasing in the detection rules and offered clearer alternatives (ref, ref, ref, ref).
How to properly debug AI Personas?
A brief yet vital discussion focused on debugging AI personas where Falco inquired about the LLM/provider specifics (ref), and Overgrow responded with the details of his OpenAI setup (ref).
This week on meta.discourse.org (Feb 17–Feb 24, 2025) the AI category has been buzzing with troubleshooting, experimentation, and feature enhancements. Users such as sam, Falco, hendersj, and others have been busy addressing critical bugs, testing integrations with local AI models like Ollama, and exploring new capabilities around PDF processing, search, and API interactions. For example, a serious page display bug raised by shannon1024 quickly attracted responses from Arkshine and sam, while robust troubleshooting in Getting discourse ai to work with ollama locally has yielded progress despite SSRF and streaming configuration challenges. Constant activity in threads discussing the performance of our AI forum helper, hosted model cost clarifications, and even Google Search API integration tests underscores a community deeply invested in making Discourse AI even better. Read on for a more detailed breakdown of the interesting topics and user activities from the past week.
Why is my AI forum helper struggling to answer questions?
In this discussion, sam breaks down common misconceptions about AI performance and walks through the inner workings of the RAG system—clarifying why even flagship models can sometimes underdeliver.
Cost of CDCK Hosted models xeraa and Falco discussed the experimental availability of hosted LLMs and embeddings, emphasizing that these features are currently free aside from fair use constraints. See follow-ups in post 2, post 3, and post 4.
PDF Support and Upload Functionality
Two threads have taken center stage in this area. In PDF support in Discourse AI, sam outlines capabilities for both basic text extraction and enhanced LLM-powered processing. This discussion continues with user excitement in Upload and discuss pdfs in composer, where MachineScholar highlights UI challenges and delayed indexing issues.
shannon1024
As the reporter of a critical page display bug, shannon1024 sparked a chain reaction of fixes and confirmations across multiple posts in that thread.
Overview
This week on meta.discourse.org the community’s AI activity has been energetic and multifaceted. Discussions have spanned innovative feature requests and detailed technical troubleshooting. A central theme has been the need to intelligently automate responses – as introduced in the Auto responder using AI discussion by Saif – with follow‐up examples from EricGT (post 2) and creative suggestions by hel_Sinki (post 3). Simultaneously, practical challenges such as enabling PDF support for AI analysis were examined in PDF support in Discourse AI with details by hameedacpa (post 7) and clarifications from sam (post 8). Meanwhile, integration issues with Claude 3.7 were dissected in the Error using Claude 3.7 Sonnet with Discourse AI plugin thread—fueling technical troubleshooting by emansilla (post 1) and iterative fixes by Falco (post 2). Other threads on LLM settings, experimental search results, and connection resets further enriched the dialogue. Collectively, these discussions underline the community’s commitment to exploring dynamic AI integrations that not only automate tasks but also enhance user experience across Discourse’s features.
PDF support in Discourse AI (#SiteManagement, how-to, ai)
• Discussions centered on testing PDF upload and analysis were led by hameedacpa (post 7) with clarifications and follow-ups by sam (post 8 and post 10), while Falco provided configuration insights (post 12).
Error using Claude 3.7 Sonnet with Discourse AI plugin (Bug, ai, ai-bot)
• Technical hurdles emerged when emansilla reported integration errors (post 1), provoking responses and troubleshooting steps from Falco (posts 2, post 4, and post 6).
Discourse AI - AI bot (#SiteManagement, ai-bot)
• A demonstration by MarkDoerr (post 150) highlighted use cases that drive the AI bot’s interactions.
Discourse AI causing new SSL and Connection Reset by Peer errors (Bug, ai)
• Issues with SSL and connectivity were investigated by oznyet (post 7 and post 9) with remedial updates from sam (posts 8 and 10).
Question about release note for experimental search results (General, ai)
• sam explained fine-tuning efforts (post 3), while Jagster shared community concerns (post 4).
Discourse AI - Large Language Model (LLM) settings page (#SiteManagement, how-to, ai)
• sam reminded users of supported configurations via the LLM settings (post 5).
How do you use Discourse AI? Tell us and make it even better! (Feature, feedback, ai)
• A feedback thread encouraged sharing of use cases, with contributions from Bhack (post 22 and post 24) and an insightful comment by Saif (post 23).
What LLM to use for Discourse AI? (#SiteManagement, how-to, ai)
• In response to user queries, Saif confirmed support for new models (post 5).
Error in any AI Tool with no parameters, e.g. “tags” (Bug, ai)
• sam detailed an error scenario associated with tool parameters (post 6).
Conversational AI Search coming to Discourse AI
Users debated ways to streamline AI search responses—suggesting interactive triggers like a clickable Ask AI button to reduce wait times and enhance conversation continuity.
Rules Surrounding Writing Topics using AI
A spirited discussion emerged about the ethics and quality of fully AI-generated topics, with community members weighing in on taste, accuracy, and moderation practices.
Rebranding the ai spam detection account
Community members clarified that rebranding the out-of-the-box AI spam detection account is safe, as internal operations depend on a fixed user ID.
Discourse AI - Large Language Model (LLM) settings page
This topic hosted technical inquiries about LLM selection, configuration, and cost/quality balance, generating discussion on optimal setups for self-hosted and hosted instances.
Using AI To Tag And Categorize Forum Posts
Members shared early experiments with automating topic categorization and tagging, discussing custom tool integrations and the nuances of silent responders.
Getting a lot of no results for semantic search
Users explored challenges with semantic search, comparing new AI conversational approaches with legacy toggled keyword searches and testing alternative configurations.
Gemini ai bot to draw picture in chat
The creative potential of integrating Gemini’s image generation capabilities into live chat was introduced, sparking excitement about future multimodal AI features.
Discourse AI spam detection replaces Akismet plugin
A crucial announcement led to discussions that clarified the scope of the changes for hosted customers versus self-hosters, as well as questions about language and performance.
Why does the AI model stop responding
Troubleshooting posts examined intermittent model timeouts, with participants sharing logs and suggestions to address the occasional failure in prolonged conversations.
This week on Meta Discourse, the community dove deep into innovations at the intersection of AI and forum management. Members explored experimental implementations such as sam's work on Experiments with AI based moderation on Discourse Meta – where discussions ranged from leveraging Gemini Flash 2.0 (post #7) to addressing false positives (post #3). At the same time, the conversation around new interaction modalities was vibrant, with topics like Is there a way to chat with Discobot? generating feedback from trusktr, tobiaseigen, and Canapin – even prompting interface tweaks such as the redirection fix by Lilly (post #10).
Experiments with AI based moderation on Discourse Meta
Discussions highlighted the promising evolution of AI for moderation, detailing experiments with Gemini Flash 2.0 (post #7) and updates ensuring that only public posts are scanned (post #6).
Is there a way to chat with Discobot?
Users explored the idea of interacting with Discobot in a chat format, with suggestions to use ask.discourse.com (post #3) and commentary on domain redirects (post #9).
What’s the cheapest/best AI to use for AI Spam?
The community weighed cost against performance by comparing models like GPT-4o-mini, Claude 3.5 Haiku, and Gemini 2.0 Flash – with hands-on recommendations from users such as trusktr (post #1) and NateDhaliwal (post #4).
FYI: Cloudflare AI Labyrinth
An insightful share on Cloudflare’s strategy to trap misbehaving bots using AI-generated mazes sparked curious reactions and humorous takes (post #1, post #3).
Hide persona from composer dropdown
A user-friendly debate was held on how to prevent personas from cluttering the composer dropdown, with suggestions for a dedicated “Hide from composer” option (post #1, post #3).
What’s the best way to add image descriptions?
Accessibility improvements were discussed with reference to current markdown capabilities and potential UI enhancements for image ALT text (post #17, post #19).
Setting to manually close AI Helper’s popup modal
With feedback about the finicky nature of the AI Helper modal, a feature request emerged to add a manual close option for enhanced usability (post #1).
Conversational AI Search coming to Discourse AI
Users provided feedback on the new search summary interface, discussing minor UI glitches like missable “More…” text and unexpected closures (post #10).
How to properly debug AI Personas?
Troubleshooting steps and questions about debugging inconsistent AI persona responses spurred a helpful discussion among experienced users (post #4).
Enrich API Calls of AI Plugin?
An enterprise user raised ideas on enriching outgoing API requests with proper authentication headers, opening a conversation on integrating internal auth endpoints with custom tooling (post #1).
This week on meta.discourse.org the community passionately discussed several AI topics, driving the conversation forward on configuration issues, user experience, and new feature experiments. Members collaborated on troubleshooting critical bugs in AI tools—ranging from errors when creating new AI personas to refining popup modal behaviors—while also sharing insights about LLM settings, sentiment analysis, image captioning, embeddings, and even experiments in AI moderation. These discussions, spread across multiple topics such as Unable to create new Personas (post 1), Setting to manually close AI Helper’s popup modal (post 3), LLM settings page (post 10) and Discourse AI - Sentiment (post 41), underscore the shared passion and technical know‐how applied by our members. The blend of bug reports, configuration tweaks, and creative ideas has paved the way for improvements to our AI functionality—ensuring that both self-hosted and managed services continue to run smoothly. For instance, fixes announced by joffreyjaffeux in Unable to create new Personas (post 6) and keegan in Setting to manually close AI Helper’s popup modal (post 3) highlight our concerted effort toward continuous improvement. Below is a detailed roundup of the most interesting topics and user activities from the past 7 days.
Interesting Topics
Unable to create new Personas (Bug, ai, ai-bot) MachineScholar kicked off this discussion by reporting an error on attempting to create new AI personas (post 1). sam joined with troubleshooting suggestions (post 2 and post 4), while further investigation by MachineScholar confirmed the issue (post 3). Later, joffreyjaffeux provided a fix by referencing the upcoming patch (post 6) and MachineScholar confirmed resolution (post 7).
Setting to manually close AI Helper’s popup modal (Bug, ai, ai-helper)
When users expressed the need for more control over AI Helper’s popup modal, keegan swiftly responded by moving the request to Bug and subsequently pushed out a fix (post 3). MachineScholar later confirmed that the issue was resolved (post 5).
Discourse AI - Large Language Model (LLM) settings page (#SiteManagement, how-to, ai)
In this thread, AquaL1te shared their experience with setting up the LLM settings page, noting a quirk with the default model name (post 10 and post 12). Saif contributed by asking for clarifications regarding model identifier discrepancies (post 11 and post 13).
Discourse AI - Sentiment (#SiteManagement, how-to, ai-sentiment, content)
This topic saw RBoy raise concerns about the cessation of the sentiment process since January 2025, prompting discussions regarding configuration changes (post 41). Falco responded with an explanation about the changes in sentiment server deployments (post 42).
Experiments with AI based moderation on Discourse Meta (Community, moderation, ai)
Creative ideas were floated by RGJ, who proposed that an AI moderation bot could intermittently signal that a post required no action—injecting an element of transparency into automated moderation (post 10).
A setting for AI-enabled image title default value (Support, ai) fbpbdmin started this discussion by suggesting that AI-enabled image titles should be off by default in order to save tokens (post 1 and post 4). keegan provided configuration guidance on the available settings (post 3).
Self-Hosting Sentiment and Emotion for DiscourseAI (#Self-Hosting, ai, ai-sentiment)
In a discussion aimed at self-hosters, RBoy raised important questions about running local instances for sentiment analysis, seeking advice on resource requirements and integration techniques (post 4).
Discourse AI - Embeddings (#SiteManagement, ai, ai-search, related-topics)
Lastly, Saif prompted an update in the embeddings discussion to ensure that the feature stays aligned with the platform’s evolving needs (post 37).
Saif
• Engaged actively in the LLM settings discussion with clarifications in post 11 and post 13, and later contributed to enhancing the AI Embeddings feature in post 37.
This week on meta.discourse.org the AI discussion threads have been buzzing with innovative ideas and practical troubleshooting. Community members explored everything from fresh feature proposals to intricate technical issues. For example, the AI Avatar Generator thread saw tpetrov kick off a discussion with ideas for generating personalized avatars (tpetrov post 1, sam post 2, tpetrov post 3), while the lively Ways to Add Knowledge to My Persona discussion featured a deep dive into enabling search and read tools for AI support personas – with contributions from Angela_MRS, Falco, and pfaffman (Angela_MRS post 1, Falco post 2, Angela_MRS post 3, Falco post 4, Angela_MRS post 5, Falco post 6, Angela_MRS post 7, Angela_MRS post 8, pfaffman post 9, plus an extra insight about private categories (Allow Bot on Private Categories)) – ensuring that even “backstage” AI teaching ideas were on the table. Other hot topics included improvements to AI‐based moderation (sam post 11) and extensive discussions on the customization and troubleshooting of the Discourse AI Bot (jorge-gbs post 152 through EricGT post 158). Meanwhile, threads on automation—such as the ability to send AI summaries to groups (sam post 2, jordan-violet post 3)—and resource management via daily AI token limits (sam post 13) highlighted both the potential and the challenges of integrating AI in community workflows. Finally, troubleshooting posts like those on the Invalid Request Error in Google Flash (BrianC post 1 through sam post 7) and API enhancements in the combined discussion on enriching API calls and Google Search integration (sam post 3 in 357898, jorge-gbs post 11 in 307107, sam post 12 in 307107) underscore the collaborative drive to refine Discourse’s AI capabilities.
Interesting Topics
AI Avatar Generator (ai, Feature): tpetrov initiated a discussion to replace static letter avatars with a dynamic, AI‐generated version. The idea was expanded by sam with technical insights (tpetrov post 1, sam post 2, tpetrov post 3).
Experiments with AI-Based Moderation on Discourse Meta (moderation, ai, Community):
A critical update by sam detailed refinements in context handling to prevent irrelevant images, helping to improve the content moderation workflow (sam post 11).
Moderation Tool for Formatting Code with AI (ai, Feature):
In an inventive proposal, sam suggested using a triage system coupled with custom tools to automatically reformat code, reducing the risk of post-destruction (sam post 11 in 347980).
Limit the Number of AI Tokens a User Can Use in a Day? (ai, completed, Feature):
Addressing resource management, sam confirmed that users now receive a notification once their daily AI token limit is hit (sam post 13 in 330500).
AI + Automation Governance: Orchestrating Independent AI Triage Scripts (ai, Support): sam outlined a workable workflow for integrating custom triage tools into the AI moderation process, signaling promising automation updates (sam post 13 in 350716).
Below is a detailed overview of this week’s vibrant discussions and developments around Discourse’s AI features and integrations from 2025-04-07 to 2025-04-14.
Experiments with AI based moderation on Discourse Meta (Community, moderation, ai)
• An experimental approach was shared by sam (post 14) to help moderators guide AI behavior.
Discourse AI - Large Language Model (LLM) settings page (#site-management, how-to, ai)
• Configuration queries were raised by jrgong (post 14), with corrective advice from Falco (post 15).
Discourse AI - Semantic Search & Summarization Not Working for our configuration (Support, ai, ai-summarize)
• Steve_John reported issues (post 1) that were probed further by sam (post 3) and elaborated with troubleshooting suggestions (post 4, post 5).
Activity
wotography led the conversation on privacy, kicking off post 1 in the Concerns over personal privacy with the AI plugin thread, and later contributed further in post 7 and post 10.
Each contribution – from detailed technical analysis to thoughtful user inquiries – has been crucial in shaping the future roadmap of Discourse AI integrations. Community members have built on each other’s ideas by referring to earlier posts like this privacy note or the troubleshooting tip, ensuring a robust and collaborative dialogue.
Thanks for reading, and I’ll see you again next week!
