# Discourse AI - Summarize
**URL:** https://meta.discourse.org/t/discourse-ai-summarize/262711
**Category:** Site Management
**Tags:** how-to, ai, ai-summarize
**Created:** [April 24, 2023, 7:40pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711 "2023-04-24T19:40:12Z")
**Posts on this page:** 20
**Page:** 1
### Author:  [@Discourse](https://meta.discourse.org/u/Discourse)
#### Post date: [April 24, 2023, 7:40pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/1 "2023-04-24T19:40:12Z")
> š This topic covers the configuration of the Summarize feature of the [Discourse AI](https://meta.discourse.org/t/discourse-ai/259214?slient=true) plugin.
>
> š Required user level: Administrator
Summarize topics and chat channels for a quick recap. Use it in mega topics and large discussions to figure out whatās happening.
## 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 generated topics
- Regenerate older summaries
- View summary date and AI model used
## Enabling Summarize
### Prerequisites
You must configure at least _one_ Large Language Model (LLM) from a provider.
To get started you can configure them through the [Discourse AI - Large Language Model (LLM) settings page](https://meta.discourse.org/t/discourse-ai-large-language-model-llm-settings-page/319903).
- [OpenAI](https://meta.discourse.org/t/discourse-ai-openai-configure-api-keys/280783)
- [Anthropic](https://meta.discourse.org/t/discourse-ai-anthropic-configure-api-key/280936)
- [Azure OpenAI](https://meta.discourse.org/t/discourse-ai-azure-openai-configure-api-keys/282442)
- AWS Bedrock with Anthropic access
- [Google Gemini](https://meta.discourse.org/t/configure-api-keys-for-google-gemini/289556)
- [Self-Hosting an OpenSource LLM](https://meta.discourse.org/t/self-hosting-an-opensource-llm-for-discourseai/290923)
### Configuration
1. Go to `Admin` ā `Plugins` ā `AI` ā `Features` and make sure the plugin is enabled (`discourse ai enabled`)
2. Select the AI Agent to use for summarization through `ai summarization agent` (a default Summarizer agent is pre-configured)
3. Checkmark `ai summarization enabled` to enable Summarize
4. Configure which groups of users can generate and view summaries by editing the allowed groups on the Summarizer agent (under `Admin` ā `Plugins` ā `AI` ā `Agents`)
5. (Optional) Enable private message (PM) summaries for specific user groups through `ai pm summarization allowed groups`
## Technical FAQ
**Does Summarize cache results?**
- Summarize does cache results and even makes them available for all users outside of selected user groups.
## Caveats
- Summarize outputs may not be 100% accurate, so make sure to check any output carefully
- LLM calls can be expensive. We recommend enabling Summarize for specific user groups to help control costs
> Last edited by @southpaw 2026-09-16T20:24:25Z
>
> Last checked by @hugh 2024-08-06T05:45:39Z
>
> > **Check document**
> >
> > Perform check on document:
---
### Author:  [@EricGT](https://meta.discourse.org/u/EricGT)
#### Post date: [July 15, 2023, 12:36pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/18 "2023-07-15T12:36:31Z")
Question
Summarization
Today summarization was activated on the OpenAI forum. In discussing it with one user found out that for moderators we see `Summarize this topic` with much less content than other users.
As moderators we get the `Summarize this topic` option
- For forum - one topic post followed by one reply post will show option
- For direct message - one topic post will show option
Tried to find documentation on this and found none. Did I miss something?
* * *
* * *
Feedback on summarization
> OMG! Thatās perfect!
>
> We need summarization, and we need it now!
([ref](https://community.openai.com/t/it-looks-like-gpt-4-32k-is-rolling-out/194615/200?u=ericgt))
The summarization topic noted is just above that reply, [here](https://community.openai.com/t/it-looks-like-gpt-4-32k-is-rolling-out/194615/199?u=ericgt).
* * *
* * *
Suggestion
Automatically drop a summary reply into topics with many replies and many views, especially discussion related topics.
Obviously there needs to be some settings so that it can be customized as to how often and which topics, which may need AI to select.
Also identify the summary repy(s) in such a way that the summarization code does not use any summary reply as that could skew the relevance of info for following summary(s).
* * *
Over the past few days as a moderator I have been hand picking topics and adding Discourse AI summaries as replies. The feedback while limited has been positive.
[Examples](https://community.openai.com/search?q=%22Summary%20created%20by%20AI.%22%20%40EricGT)
* * *
* * *
Crazy suggestion
Create Discourse news site
* * *
Every day I visit many sites to get updates. Some of those sites provide news articles about the latest AI trends. Many of those news articles are really just many one liners of facts. One [news article](https://www.businessinsider.com/openai-gpt4-ai-model-got-lazier-dumber-chatgpt-2023-7?ref=emergentmind) I read this morning was pulled from the OpenAI Discourse forum and Discord forums. In using the summarization for that topic noticed that what was [generated](https://community.openai.com/t/experiencing-decreased-performance-with-chatgpt-4/234269/176?u=ericgt) was just as good or better than the news article.
Now many of us know about [Hacker News](https://news.ycombinator.com/).
So why not have Discourse news which pulls news based on Discourse sites that agree to have content made available on the Discourse news site. Obviously there would have to be opt-in options for each site, user and topic but you never know, could become a new source of info for [The pile](https://en.wikipedia.org/wiki/The_Pile_(dataset)) and if it also done with the proper approval would be a benefit to many.
* * *
* * *
Feedback
Created a summary for this topic but did not publicly post
> **[Code interpreter is turned off, any idea?](https://community.openai.com/t/code-interpreter-is-turned-off-any-idea/299217)**
>
> The code interpreter is offline, does anyone know whatās going on?
The summary was created after this post (number 5)
> **[Code interpreter is turned off, any idea?](https://community.openai.com/t/code-interpreter-is-turned-off-any-idea/299217/5)**
>
> Resolved - Code Interpreter is operational. Seems Matt of the OpenAI staff is active on the Discord channel (link)
Summary
> User jy.genre reported the code interpreter being offline with an uploaded image. EricGT responded, saying it was live for them and produced code. They also shared images of the Discourse forum and the OpenAI status page showing a maintenance message, which they had not seen before. They also shared updated maintenance messages. Markanthonykoop also reported seeing a maintenance popup and had experienced the service looping previously. Magejosh confirmed that they couldnāt open new code interpreter chats but could upload files to previously opened chats. They noticed more frequent repeating mistakes and instructions being forgotten, but presumed this was due to maintenance. EricGT shared another maintenance update from the OpenAI status page.
Two observations
- The time or sequence of the events is important and missing from the summary
- There is a lot of critical information in the images that is not being extracted for the summary, perhaps OCR can be used.
---
### Author:  [@EricGT](https://meta.discourse.org/u/EricGT)
#### Post date: [July 20, 2023, 5:44am UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/24 "2023-07-20T05:44:55Z")
The original topic is in the Lounge category on OpenAI Discourse which most can not access.
Posted here for Roman and Falco to see and to also save Sam the work or relaying this; Sam is doing great staying in touch on summarization at OpenAI.
> [@How to easily edit my site remotely?](https://meta.discourse.org/t/how-to-easily-edit-my-site-remotely/300940/19):
>
> āreboot this very noisy topic with a new topic that is a lot clearerā
I like the reboot thought.
Also consider similar topics that should be merged into one, e.g
> **[GPT has been severely downgraded](https://community.openai.com/t/gpt-has-been-severely-downgraded/260152)**
>
> Many of us using the GPT Playground for activities ranging from low-level coding and beyond have noticed a significant drop in performance with GPT-4. While OpenAI hasnāt officially acknowledged this change, itās abundantly clear to those of us...
> **[Has There Been A Recent Decrease In GPT-4 Quality?](https://community.openai.com/t/has-there-been-a-recent-decrease-in-gpt-4-quality/207392)**
>
> Iāve been using Chat GPT for quite a while now, and Iāve been a GPT Plus user since GPT-4 was released. Over the past few days, it seems like GPT-4 is struggling to do things it did well previously. I use GPT-4 to augment long-form content...
> **[Experiencing Decreased Performance with ChatGPT-4](https://community.openai.com/t/experiencing-decreased-performance-with-chatgpt-4/234269)**
>
> Has OpenAI provided an explanation for the very clear degradation of ChatGPTās performance with the GPT4 model? Since the last two updates, it is significantly faster (which is fundamentally useless when we are capped at 25 messages in 3 hoursā¦),...
> **[Terrible performance degradation GPT Plus](https://community.openai.com/t/terrible-performance-degradation-gpt-plus/267918)**
>
> Guys, listen, it is serious, GPT-4 performance sucks. After 12th May release, (now 24th May) something really went wrong. In all aspects. GPT-4 now makes stupid mistakes, plus it forgets every 5th or 6th conversation. Wake up ā¦ If this...
> **[GPt-4 is underwhelming now](https://community.openai.com/t/gpt-4-is-underwhelming-now/304936)**
>
> GPt-4 officially underwhelming now.
---
### Author:  [@JammyDodger](https://meta.discourse.org/u/JammyDodger)
#### Post date: [September 7, 2023, 9:53am UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/25 "2023-09-07T09:53:46Z")
14 posts were split to a new topic: [Add more language support for AI summaries](https://meta.discourse.org/t/add-more-language-support-for-ai-summaries/278151)
---
### Author:  [@Hifihedgehog](https://meta.discourse.org/u/Hifihedgehog)
#### Post date: [September 7, 2023, 5:14pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/40 "2023-09-07T17:14:41Z")
Is there a way to clear cached summaries from using other LLMs? I ask since one LLM gave me this beauty of a summary, so I have since switched to another, but I am not certain how to remove this old summary from the forum system.
---
### Author:  [@Falco](https://meta.discourse.org/u/Falco)
#### Post date: [September 7, 2023, 5:17pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/41 "2023-09-07T17:17:30Z")
Currently staff can regenerate summaries after 1h. Other way is using the Ruby console to delete a specific cached summary from the DB.
---
### Author:  [@Falco](https://meta.discourse.org/u/Falco)
#### Post date: [September 7, 2023, 5:26pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/42 "2023-09-07T17:26:05Z")
By the way, we didnāt find the existing summarization models good enough for Discourse, and have switched to using LLMs for this. If you have a server with enough GPU VRAM, running a Llama2 based LLM will get you great results for #ai-summarization and #ai-helper. I have updated [Discourse AI - Self-Hosted Guide](https://meta.discourse.org/t/discourse-ai-self-hosted-guide/259598) with basic instructions about how to run said LLM.
---
### Author:  [@pfaffman](https://meta.discourse.org/u/pfaffman)
#### Post date: [September 8, 2023, 12:15am UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/43 "2023-09-08T00:15:26Z")
> [@Falco](#):
>
> If you have a server with enough GPU VRAM
Iām annoyed when people ask this, but can you give a hint what āenoughā is?
Maybe ASUS Dual NVIDIA GeForce RTX 3060 V2 OC Edition 12GB GDDR6 Gaming Graphics Card (PCIe 4.0, 12GB GDDR6 Memory, HDMI 2.1, DisplayPort 1.4a, 2-Slot, Axial-tech Fan Design, 0dB Technology) [Amazon.com: ASUS NVIDIA GeForce RTXā¢ 3050 LP BRK OC Edition 6GB GDDR6 (IP5X dust Resistance, Dual Ball Bearings, Stainless Steel Bracket, PCIe 4.0, DLSS 3, HDMI 2.1, DisplayPort 1.4a, DVI-D), 3 Year Warranty : Electronics](https://a.co/d/8LSAd8A)
---
### Author:  [@Falco](https://meta.discourse.org/u/Falco)
#### Post date: [September 8, 2023, 12:58am UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/44 "2023-09-08T00:58:28Z")
> [@pfaffman](#):
>
> Iām annoyed when people ask this, but can you give a hint what āenoughā is?
You should take a look at the edit I linked above, it has just that!
But the longer version of it is that quantization allows you to trade-off quality and speed for running on lower specs. If you want to run the best models, without quantization you need around 160GB VRAM. If you accept worse quality, speed, latency, etc you can use 5% of that.
We have āgood enoughā results running the model I use as an example there, which barely runs in a machine with 96GB of VRAM.
---
### Author:  [@Hifihedgehog](https://meta.discourse.org/u/Hifihedgehog)
#### Post date: [September 11, 2023, 3:22am UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/45 "2023-09-11T03:22:42Z")
> [@Falco](#):
>
> Currently staff can regenerate summaries after 1h.
I may be missing it so pardon me if I am, but where is the option to force a fresh regeneration for staff?
---
### Author:  [@Hifihedgehog](https://meta.discourse.org/u/Hifihedgehog)
#### Post date: [October 10, 2023, 4:57pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/46 "2023-10-10T16:57:39Z")
> [@Falco](#):
>
> Currently staff can regenerate summaries after 1h. Other way is using the Ruby console to delete a specific cached summary from the DB.
Sorry for the bump, but I havenāt been able to find the method for staff to regenerate summaries after 1 hour. Does this 1 hour window require that reply has been made to the topic, is there some secret UI combination that has to be performed to access this, etc.?
---
### Author:  [@swong](https://meta.discourse.org/u/swong)
#### Post date: [October 31, 2023, 6:39pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/47 "2023-10-31T18:39:48Z")
Is there any way to use the `Anthropic's claude-2` summarization strategy, using AWS Bedrock with Anthropic access (AI Chat Bot has this option)?
Currently, either Anthropic model can only be selected if `ai_anthropic_api_key` is configured.
---
### Author:  [@Falco](https://meta.discourse.org/u/Falco)
#### Post date: [October 31, 2023, 7:34pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/48 "2023-10-31T19:34:34Z")
Yes, you need to
- Set `ai_anthropic_api_key` to āaā (to bypass the validation)
- Fill your bedrock credentials
- Select claude 2 as the summarization model
We have a planned overhaul of the way models / inference / APIs can be selected in the UI, but this will work for you in the meanwhile.
---
### Author:  [@EricGT](https://meta.discourse.org/u/EricGT)
#### Post date: [November 24, 2023, 12:19pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/49 "2023-11-24T12:19:48Z")
Feedback.
For the most part the summary is working fine.
For links to images in the post the links are not correct.
> **[OpenAI Dev Day 2023 Live Reactions](https://community.openai.com/t/openai-dev-day-2023-live-reactions/475167/78?u=ericgt)**
>
> Summary created by AI. The OpenAI Dev Day 2023 live reactions started with elmstedt indicating that the Developer Day presentations had commenced and an encouragement to build something cool using the new toys introduced. First, RonaldGRuckus...
e.g.
`Following a few photographs shared by [Foxabilo](/t/-/475167/13; /t/-/475167/24; /t/-/475167/34; /t/-/475167/37)`
---
### Author:  [@Jagster](https://meta.discourse.org/u/Jagster)
#### Post date: [November 24, 2023, 12:28pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/50 "2023-11-24T12:28:39Z")
This is part of good olā question of languagesā¦ but has someone an idea why sometimes summarization respects language of topic and sometimes it uses english? It looks like it happends totally randomly.
And actual question could be is there a system prompt where we can suggest used language or does it come from summarize tool?
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [November 27, 2023, 3:34am UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/51 "2023-11-27T03:34:14Z")
> [@Jagster](#):
>
> And actual question could be is there a system prompt where we can suggest used language or does it come from summarize tool?
This is hard coded internally atm, but we plan to allow flexibility here.
Some people like longer summaries, others like shorter ones ā¦ etcā¦
---
### Author:  [@RBoy](https://meta.discourse.org/u/RBoy)
#### Post date: [February 13, 2024, 4:40pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/54 "2024-02-13T16:40:55Z")
> [@Falco](#):
>
> Available strategies include:
>
> - A collection of open-source models from HuggingFace: bart-large-cnn-samsum, flan-t5-base-samsum, and long-t5-tglobal-base-16384-book-summary.
> - OpenAI ChatGPT, using either gpt-3.5-turbo or gpt-4.
> - Anthropic Claudeās V1 and V2 models.
This text may need an updated as I believe it supports Gemini also.
Additionally, feedback on the settings UI, was there any particular reason that the `summarization_strategy` and `custom summarization allowed groups` settings were moved to the `Others` page instead of being with the `Discourse AI` page along with the rest of the AI settings? It look a while to find this and itās getting lost among a set of unrelated settings.
---
### Author:  [@RBoy](https://meta.discourse.org/u/RBoy)
#### Post date: [February 13, 2024, 8:02pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/55 "2024-02-13T20:02:08Z")
How does one cycle this feature on/off? Iām not seeing the Summarize button for topics with the required number of posts and also seeing messages in error logs so I want to try to turn it off and then on again.
---
### Author:  [@Shauny](https://meta.discourse.org/u/Shauny)
#### Post date: [April 1, 2024, 9:29am UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/56 "2024-04-01T09:29:07Z")
How can we completely disable Summarize with AI?
The quality is really bad and I just find the whole button pointless - Iād like to remove it.
---
### Author:  [@Falco](https://meta.discourse.org/u/Falco)
#### Post date: [April 1, 2024, 12:49pm UTC](https://meta.discourse.org/t/discourse-ai-summarize/262711/57 "2024-04-01T12:49:23Z")
> [@Shauny](#):
>
> How can we completely disable Summarize with AI?
The same way you enabled it with the `summarization strategy` setting. To disable it you can reset it to the default site setting value.
> [@Shauny](#):
>
> The quality is really bad
What model are you using? We found the Claude 3 Opus performs best at this task.
[Next page](https://meta.discourse.org/t/discourse-ai-summarize/262711.md?page=2)