support-chat

Hi E Salinas @drjosegregorio,

Does our Talkyard instance/version support OIDC/OAuth2 authentication? If , how [... ]

Yes. A bit higher up on the login settings page,
(hmm would be good if this was simpler to find, not "hidden" in the middle of a long page?)
you should find this title:

Ways to sign up

and just below the title, a checkbox:   Custom OIDC or OAuth2

If you tick the checkbox, a button appears:
Configure identity Providers (IDPs) ...

(However, if you have enabled Enable Single Sign-On (SSO), then the OIDC settings are hidden, since can't use both at the same time.)

Click the button, and an OIDC configuration field appears, where you can copy-paste & edit OIDC configuration.

Read here about how to configure: How to enable OIDC login

Haven't tried to connect Talkyard to Supabase, but if they use standard OIDC, it should just work.

Hello! Is there an api for getting embedded comments for a certain page in json format? List Posts endpoint returns everything all together and filters are not implemented. Get endpoint returns only likes count, and if i want detailed information i need to use api key which doesn't make sense for the embedded comments case. Is there anything else?

Hi @Daria, Listing comments per page — yes, not yet implemented, but should be a somewhat quick fix. I can write more in a few days / later this week.

Is there anything else?

Maybe depends on what you want to do :- ) What's your use case, if I can ask?

(I'm guessing, though, that the best approach is to implement the get-comments-per-page functionality.)

API docs page: https://docs.talkyard.io/api/v0#tag/default/post/-/v0/list

Relevant source code in class ListController: (aborts early, since not implemented)

    throwUnimplementedIf(anyFilter.isDefined && (listWhat == Posts || listWhat == Events),
          "TyE603MRD4", "No filters implemented for posts and events")

I am using the API to create posts and comments. Is it possible to increase the size of the POST messages? I am guessing this is something in the backend webserver talkyard is running?

I get this error, when trying to create a page using the API (thru a POST call in Python or curl)

❌ An error occurred: 413 Client Error: Payload Too Large for url: https://archprisms.talkyard.net/-/v0/do

server_origin = "https://archprisms.talkyard.net"  # e.g., "https://example.com"
secret_b64 = "REMOVED"
username = "karu"
page_ref_id = sys.argv[1]
cat_ref_id = "rid:category_questions"

# --- Define your multiline string with special characters ---
# Using a triple-quoted string makes this easy and readable.
body_content = """
Review of the paper "LightML: A Photonic Accelerator for Efficient General Purpose Machine Learning," written from the perspective of "The Guardian."

### Review Form

**Summary**

This paper introduces LightML, a photonic co-processor architecture designed for general-purpose machine learning acceleration. The authors claim to present the first complete "system-level" photonic crossbar design, including a dedicated memory and buffer architecture. The core of the accelerator is a photonic crossbar that performs matrix-matrix multiplication (MMM) using coherent light interference. The paper also proposes methods for implementing other necessary ML functions, such as element-wise operations and non-linear activation functions, directly on the photonic hardware. The headline claims are a peak performance of 325 TOP/s at only 3 watts of power and significant latency improvements (up to 4x) over an NVIDIA A100 GPU for certain models.

**Strengths**

The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:
The paper is well-written and addresses a compelling long-term research direction. The core strengths are:

"""

# --- Prepare the request ---
url = f"{server_origin}/-/v0/do"

headers = {
    "Authorization": f"Basic {secret_b64}",
    "Content-Type": "application/json"
}

# The payload is a Python dictionary, which is easy to construct.
payload = {
    "doActions": [{
        "asWho": f"username:{username}",
        "doWhat": "CreatePage",
        "doHow": {
            "refId": page_ref_id,
            "pageType": "Discussion",
            "inCategory": cat_ref_id,
            "title": "Bird vs Owl (with details)",
            "bodySrc": body_content, # Your multiline string is inserted here
            "bodyFmt": "CommonMark"
        }
    }]
}

# --- Send the request ---
try:
    # The `json` parameter automatically converts the dict to a JSON string
    print(url)
    print(headers)
    print(payload)
    response = requests.post(url, headers=headers, json=payload)

    # Raise an exception for bad status codes (4xx or 5xx)
    response.raise_for_status()

    print("✅ Success!")
    print(f"Status Code: {response.status_code}")
    # Print the JSON response from the server
    print("Response JSON:", response.json())

except requests.exceptions.RequestException as e:
    print(f"❌ An error occurred: {e}")

@karu It seems there's a just 2000 bytes limit!

  def apiV0_do(): Action[JsValue] = ApiSecretPostJsonAction(  // [PUB_API]
          // For now, may do just a few things. [do_api_limits]
          WhatApiSecret.SiteSecret, RateLimits.UpsertFew, maxBytes = 2000,
          ) { req: JsonPostRequest =>
    queryDoImpl(req, doOnly = true)
  }

(appsv/server/talkyard/server/api/QueryDoController.scala)

Is it possible to increase the size of the POST messages?

Yes, I'm about to release a new version now soon, some time week I'm guessing. I'll look into making the limits configurable. (Sorry for the temporary blocker)

using the API to create posts and comments

Ok, good idea. Mostly just out of curiosity, how large texts would you expect the AIs to generate?

Thank you!

Typical size is around 6000 characters. Would be nice to support up to 20K characters.

@KajMagnus Hi Magnus...just wondering if you had an update on the ETA for this feature.

@karu At the end of this week I think (Fri, Sat or Sun)

@karu Now I've upgraded the server — 20K chars (and more than that) is now ok.

(Not impossible that there'll be some other limit you'd run into — then please let me know)

Hello. Our community portal hosted in cloud seems to be unreachable now. Is there any outage at this moment?

same issue for us today and worked just fine last night, no changes in DNS records or anything related ... A timeout occurred Error code 524
Visit cloudflare.com for more information.
2025-10-02 14:21:51 UTC

and its back and working now :) , just to confirm this is something you fixed/restarted on TY side or this is something else ?

Started working on our end too. Also curious what was that?

wondering if this is CF related issue https://www.cloudflarestatus.com/incidents/k8wn9d8kn7tz

We're down again this morning. The domain is not responding... Might be an issue on the CF side I guess, but I can't say for sure.

yes - i can confirm it - this is not good - and its not CF issue at this point this is TYard

SSL Provisioning Delays
Incident Report for Cloudflare
Resolved
This incident has been resolved.
Posted 9 hours ago. Oct 03, 2025 - 04:01 UTC

@KajMagnus may we have your attention here please?

I guess its no go for this type of forum :) google bots see webpages not available few times each day and simply stop crawling going forward

502 Bad Gateway [TyE502BGW]

Talkyard's Nginx server cannot connect to Talkyard's application server.
Is the application server not running, or is there some network error?

Please check if the 'app' Docker container is running: 'docker-compose ps'

If it's not running:

• Start it: 'docker-compose start app'. Then wait a few seconds and reload this page.
• Or login in a Bash shell: 'docker-compose run --rm --service-ports app bash'

If it is running:

• Check the logs: 'docker-compose logs'
• Only Play's logs: 'docker-compose logs app'
• Or jump into the container: 'docker-compose exec app bash'
  ---Invalid SSL certificate Error code 526
  Visit cloudflare.com for more information.
  2025-10-03 18:10:32 UTC
  You
  Browser
  Working
  Newark
  Cloudflare
  Working
  forum.stefanybags.com
  Host
  Error
  What happened?
  The origin web server does not have a valid SSL certificate.

What can I do?
If you're a visitor of this website:
Please try again in a few minutes.

If you're the owner of this website:
The SSL certificate presented by the server did not pass validation. This could indicate an expired SSL certificate or a certificate that does not include the requested domain name. Please contact your hosting provider to ensure that an up-to-date and valid SSL certificate issued by a Certificate Authority is configured for this domain name on the origin server. Additional troubleshooting information here.

I'd still appreciate feedback from the team. Do you think it was an exception and something that got fixed already?

Hi @yurist38 and @stefanybags, sorry yes the server was malfunctioning. I had changed the amount of memory from 16 GB to 8 GB but I hadn't updated the application server, so it thought it was supposed to use a lot more memory, than what was actually available.

Just after I made the change, all seemed fine, but after some day, the app server started using more and more memory, without me noticing, and eventually it stopped working.

I thought I had configured alerts that would notify me about this, but seems that didn't work, ... need to look into that. And probably there should be a checklist for changing the amount of memory so I won't forget some other time too

(A question is: How does one remember if there's a checklist for doing something, or not. How does one remember to look at the checklist)

@yurist38 and @stefanybags, I'll send you a DM in a little while

@KajMagnus good to hear that the issue is found and fixed now. Thank you for the update, I appreciate it.

Turns out I did get an alert notification about the server being offline, but I didn't notice partly because I was doing other things, partly because I got some other & more recent notifications about possibly problematic log messages — which are always almost harmless. Now I've reduced how often notifications are sent about possibly-problematic log messages, so I'll better notice the actually important alerts

(Interesting, seems pilots sometimes have some minor problems remembering to check their checklists:

"How do I remind myself to run in-flight checklists?"
https://aviation.stackexchange.com/questions/38476/how-do-i-remind-myself-to-run-in-flight-checklists

"How do you remind yourself to do your checklists?"
https://www.pilotsofamerica.com/community/threads/how-do-you-remind-yourself-to-do-your-checklists.102704/

At least there's some seemingly good advice there)

on page https://www.talkyard.io/compare/discourse
It says Install on-premise as open source (open core)

Is this still the case? what parts are not open now?

https://github.com/debiki/talkyard-prod-one?tab=readme-ov-file#installation-instructions

step 7:

Update 2021-04-04: It's better if you use firewalld instead — it's safer: Docker can bypas ufw rules, but not firewalld rules.

The script is still using ufw

@IvanTheGeek Everything in Talkyard today is open source.

SCIM and SAML (enterprise stuff) when/if implemented in the future, would probably be closed source (or source-available). And some future enterprise compliance stuff that I don't know exactly what it is.

The script is still using ufw

In the upcoming Talkyard v1 the ufw instructions will be gone, instead, just this text:

 Install a firewall, for example firewalld. Note that ufw (another Linux firewall)
 is incompatible with Docker, see:
 https://docs.docker.com/engine/network/packet-filtering-firewalls/#docker-and-ufw.
 (If you use Google Cloud Engine: GCE already has a firewall.)

Maybe I could make those edits now directly instead, no need to wait until v1 hmm.

... Now there's some edits in this main2 work-in-progress branch, with ufw removed: https://github.com/debiki/talkyard-prod-one/blob/main2/README.md#preparations

In looking at that, seems you are changing license to MIT?

Just did a signup for the trial (you hosting). One the agreements page, the links to the TOS and Privacy policy end up being not served pages. [YES, I look over the crazy stuff like that]

@IvanTheGeek Oh you're right. Thanks for mentioning!

So, when having created a forum, and logging in as forum owner the first time, the TOS and Privacy both link to: https://www.talkyard.io/-/terms-of-use and https://www.talkyard.io/-/privacy-policy, which redirects to: https://forum.talkyard.io/-terms/of-use and https://forum.talkyard.io/-privacy/policy, both of which are 404 Not Found.

Fortunately, it's only when logging in the 1st time as owner — thereafter, when other people join the forum, the links work (they point to https://ones-own-forum/-/terms-of-use and .../-/privacy-policy).

Either I should 1) remove this dialog, when joining the forum as the forum owner. It's his/her forum, and s/he doesn't need to accept their own terms. Or 2) link to some SaaS terms instead, hmm. There's some automated emails sent later with links to the SaaS terms, you might have gotten already.

@IvanTheGeek The installation scripts repo: https://github.com/debiki/talkyard-prod-one is MIT and has always been. But it's just bash scripts and documentation. The actual Talkyard discussion forum software is: https://github.com/debiki/talkyard and is AGPL.

Thanks for the license explanation. I did not catch that. For the admin setup, I am thinking since I was doing a saas on your servers, I was expecting to find the terms and policy related to me having you host my forum, if that helps.

In signing up as my forum user, I would expect to find the forum owners (mine) tos and privacy policies.

Just feed back, hope it helps.

@IvanTheGeek Ok :- ) Now I've fixed the site owner ToS link, so it points to: https://legal.talkyard.io/talkyard-standard-saas-terms

The site owner privacy link — for now, it's: https://forum.talkyard.io/-/privacy-policy, but it's a bit messy — it's a mix of site owner and normal forum member privacy-policy :-/ Probably it'd be better if there was a new Privacy Policy for site owners owner only (& maybe site admins too).

find the terms and policy related to me having you host my forum

Yes, that was my idea too

Just feed back, hope it helps.

Yes it did / does :- )

Its nice that you have reacted so quickly!

FEEDBACK:

• the text is clear and readable
• the shortcuts are a nice touch
• also, the reference to where the terms are based off is nice
• Section 2. "When using our Services requires an account," the grammer is off somehow. Maybe either "Using our Services requires an account," or "An account is required when using our Services," or something of that nature.
• Section 2. the link (https://www.talkyard.io/-/privacy-policy) to Privacy Policy is invalid

Is there any consideration or ability to support IPv6?

Doing a whois on the IP for c1.talkyard.net (the CNAME I was told when I went to add my own domain name in admin) is showing Google Cloud. You may be using them directly or you are the customer of someone using them for infrastructure.
Either way, my cursory search about it for google is that IPv6 is included at no charge.

Two areas would need addressed:

1. The dns record for c1.talkyard.net would need a new AAAA record pointing to the assigned IPv6 address. This would be required on your side as the c1.talkyard.net domain will need to reply to A and AAAA requests. Once I set my custom domain (i.e. forum.example.com) as a CNAME, I can not add a A or AAAA record to it.
2. the server would need the IPv6 address added to its networking stack, and then whatever server is running Talkyard would need to be attached to that IPv6 address as well.

OPINION: IPv6 will never reach its full potential until as many folks that ARE able to support IPv6 ARE supporting it. As Talkyard seems to be somewhat software and technical orientated forum software (which is one of the attractions as it is more the way MY mind works), it would seem to be a good option to support the modern practices.

As I was writing a previous post, I wanted to use incline code block code. I was able to do so from knowing about already. I am more of a click guy, so I clicked the </> thinking it would do the code for the highlighted selection. It instead pull it down into a full code block with new lines. Having a logic or separate button in the editor to do inline code would be desired.

If you rather I post these things in the a topic under a category or something. Pleas let me know!

@IvanTheGeek Thanks for the feedback! Oh it's the same broken link, in another source code repo (a legal docs repo)

I haven't thought about IPv6. Actually I think it'll have to wait :- / Thanks for mentioning the areas-to-address anyway. I agree with your opinion, just that there's so much to do, so little time

About </> — hmm, if using </> for inline code, what button icon makes sense for a multi line code block button? (I asked ChatGPT but it didn't have any good ideas :- ))

If you rather I post these things in the a topic under a category or something. Pleas let me know!

Yes for new ideas / issues in the future, I'd prefer that :- )