Community Strategy Insights

The latest insights on community strategy, technology, and value by FeverBee’s founder, Richard Millington

Should You Build Your Community On Discourse?

Richard Millington
Richard Millington

Founder of FeverBee

Of every platform I review in this series, Discourse is the one I score highest on the thing communities are actually built to do: hold a good discussion. That is a 9 out of 10, the top score I give anyone. It is also one of the three platforms, alongside Higher Logic Vanilla and Gainsight, that I see Khoros communities migrating to most often.

I run it through the same nine-capability scorecard as every episode, so you can compare it like for like. Here is why it is so strong, where it falls down, and the one thing you have to get right if you move to it.

Where Khoros communities are going

I will start with the evidence for that claim, because it is not a hunch.

Look at where some of the biggest communities leaving Khoros have actually landed. Google moved its developer community and its Local Guides community onto Discourse. Shopify runs its community on it. And HubSpot have just relaunched their community on it too.

These are not small forums. As I record this, the Shopify Community on Discourse has around 1.5 million members and 1.8 million posts. Google’s developer forums have over 260,000 members and close to a million posts. HubSpot has around 500,000 members and 800,000 posts. Those are live numbers, and they are some of the largest brand communities on the planet. When they left Khoros, this is where a lot of them went.

So when I say Discourse is one of the three places Khoros communities most often land, alongside Higher Logic Vanilla and Gainsight, that is what I am basing it on. The question for you is not whether serious companies trust it. It is whether it fits what your community is for.

A product-first company

To understand Discourse, you have to understand that it has been a product-first company for its entire life, since Jeff Atwood founded it in 2013.

That came with an advantage and a disadvantage. The advantage is that they built the best forum-centric product in the market, by a distance, because the product was the obsession. The disadvantage is that for years they were behind on the things around the product: support, sales, professional services, the commercial machinery other vendors had figured out. If you wanted a slick sales process and a big account team, Discourse historically was not where you would find it.

That history still shapes the platform you would buy today. The product runs deep. Outside the product, the rest of it is catching up, and it matters if your last impression of Discourse is a few years old.

They have built a proper team, and it is a good one. They are the most technically savvy group I engage with in this space. Almost everyone I have spoken to there actually knows how to use the platform and understands community better than the sales teams I deal with at most other vendors. That is rare, and it is the precise thing people leaving Khoros say they stopped getting.

There are two honest caveats. The product-first focus still lingers a little. And they are still a relatively small team compared to the bigger organisations serving this space. But they punch well above their weight, and where it counts, the people are excellent. That product DNA also shapes where they are strongest: technically minded communities, developers and gaming especially, the places where code blocks and the deeper forum mechanics genuinely matter.

Discussions and Q&A

Discussions and Q&A. I score Discourse a 9, the highest discussion score I give any platform in this entire series.

Discourse is the benchmark other platforms get measured against. It tracks each member’s read position, so it knows where everyone left off, and it lets the community edit posts together in wiki mode. The composer handles code blocks with proper syntax highlighting. You also get real-time presence, topic timers, slow mode, whispers, custom reactions, bookmarks, and tag groups. The admin tooling underneath is excellent: you can merge, split, move, and archive topics, run bulk actions, pull up full revision history, and schedule auto-close and auto-bump. If running a serious discussion is the job, nobody does it better.

On the Q&A side, there is a plugin that gives you accepted answers and a proper question-and-answer format, and it works well. Discourse goes deeper on raw discussion craft than anything else I review.

One recent thing worth knowing: nested replies only shipped in mid-2026. So if you looked at Discourse a year ago and the flat reply structure put you off, look again.

AI and the MCP

Second, and this is where Discourse has genuinely surprised people. AI. I score it an 8.

There is a live MCP, the connection that lets external AI tools read and act inside your community through a standard interface. There are AI Agents, thirteen tools that can close a topic, edit tags, set slow mode, grant a badge, or mark a question as solved, with a human reviewing before anything goes live. There is AI search across the whole community. It writes summaries of long topics, auto-tags new posts, runs AI spam detection, and there is an AI bot that answers questions straight from your community’s content.

Two of those deserve a closer look.

First, the MCP. Plenty of platforms are now bolting on a way for AI to read your community. What makes Discourse’s different is that it does not just let an assistant read, it lets it act, through those thirteen agent tools, and it sits on a community that is already machine-readable from the ground up.

Which is the second, and smartest, thing they have done. Every page in a Discourse community is machine-readable. Add “.json” to the end of any topic’s web address and you get a clean, structured version of that thread: the question, the answers, who said what, and which reply was marked as the solution. You do not switch it on. It is just there. So when someone asks ChatGPT or Claude a question your community has already answered, Discourse is one of the formats those tools find easiest to read and to cite. They have gone further, with support for the newer llms.txt standard so you can hand AI tools a curated map of your best content.

The practical upshot: if being found inside AI answers matters to you, and it should, Discourse is unusually well set up for it. The reason this scores an 8 is that it is shipping, not roadmap. They put this stuff out month after month.

Moderation

Third strength, and it connects straight to the last one. Moderation. Also an 8.

The same AI agents carry a set of moderation-specific tools, paired with that human review queue, so you get AI doing the first pass and a person making the call. Underneath sits Discourse’s trust level system, which automatically promotes members through four tiers as they earn it, which quietly takes a lot of routine moderation load off your team. Add AI spam detection and a deep plugin ecosystem and you have one of the stronger moderation suites in the market.

This matters more than it used to. If you are a large consumer community with real safety concerns, moderation is no longer a nice-to-have. Discourse holds up well here, and it is a common reason I see enterprise teams shortlisting it.

Where it is middle of the pack

Now the capabilities where Discourse is fine, but they are not why you would choose it.

Gamification, a 6. The trust level system is an elegant hybrid: it gamifies participation and reduces moderation burden at once, because earning trust unlocks abilities. There are badges with custom triggers, and an official plugin that adds leaderboards and scoring. But it does not have the deep, native points-and-ranks reputation engine Khoros built, which is why Khoros still scores a 7 here and Discourse sits just behind.

Content and knowledge, a 5. You get the basic building blocks: a Docs plugin that gives a knowledge-base style view, wiki mode so the community can maintain articles, and a proper rich-text editor since 2025. But it is not a real knowledge platform. No native courses, no blog, no video library, and the knowledge-base experience is lighter than the tools built specifically for it. It is adequate. It is not a reason to buy.

Mobile, a 5. A strong progressive web app that feels close to native, a Discourse Hub app for iOS and Android, and the option of a custom branded build. Better than Vanilla at a 3, behind Khoros at a 6. Solid, not a selling point.

Analytics, a 5. There is built-in reporting for content and engagement, extended by plugins. One point in its favour: because Discourse is so open, it is easier to get your data out and into your own analytics tools than with most platforms. But the reporting inside the product is still a bit weak, so if your execs want board-ready dashboards out of the box, you will feel that gap and lean on the API to close it.

Customisation and depth: the thing you have to get right

Now the thing that defines the whole platform, and the single decision that makes or breaks a move to Discourse.

Discourse has some of the most advanced features and functionality out there, and with a deep plugin ecosystem plus the ability to get right into the code, the ceiling on what you can build is effectively unlimited. The critical thing, if you move to it, is that you actually take advantage of that. The whole value of Discourse is that it goes really deep in a handful of areas: discussion mechanics, moderation, AI readiness, the developer and gaming use cases. If you are inside those areas and you use the depth, almost nothing competes. If you are outside them, that depth is not doing anything for you, and the capabilities we just walked through are where you would feel short-changed.

There are two things to consider. First, a lot of that power lives in plugins and custom development, so you want a developer in the loop, or budget for Discourse’s own services team. A large enterprise team currently moving off Khoros told me they had had to install some plugins and do a little custom development, but that it was easier than the heavy-handed custom work Khoros had required. That matches what I hear repeatedly.

Second, it can be a lot. Discourse is so feature-rich it can feel overwhelming, on the front end for members and on the back end for admins. This is not a simplified forum experience. It is a deep forum experience.

So here is the failure mode I would warn you about, because I have seen it. A team buys Discourse for the depth, then treats it like a simpler packaged platform. They never put the technical resource behind it. They do not build the plugins, they do not use the deep mechanics, and they end up with something more complex than they needed and less finished than what they left. The platform did not fail them. They bought the depth and never used it. If you are not going to resource it, a more packaged platform will make you happier.

One practical note for your evaluation. You do not get a sandbox to trial your own configured instance the way the usual enterprise platforms hand you one. There is a generic demo you can poke at, but testing Discourse properly means committing to a real build. That raises the stakes on getting the decision right up front, so do the homework before you sign.

The gaps: events and ideation

Now the two weakest areas, stated plainly.

Events, a 3. Events used to be a bolt-on plugin. It is now bundled into Discourse core, with an interactive calendar, event posts with going and interested RSVPs, recurring events, timezones, even a holiday calendar. That is real, and it is well ahead of Khoros, which has nothing native and scores a 1. What it still is not is a full events platform. There is no ticketing, no registration pages, no check-in. So if meetups and webinars are the point of your community, this will not be enough, and you would want Bevy or Gradual.

Ideation, a 3. There is a Topic Voting plugin that creates vote-sorted idea categories, and real communities use it in production, but it is not a purpose-built ideation workflow. If feeding your product roadmap is the main point of your community, Gainsight is miles ahead at an 8, with idea scoring weighted by account value.

Neither is fatal for a discussion-led community. Both are dealbreakers if they are central to what your community is for. Know which you are before you sign.

Is Discourse a healthy vendor?

In the Khoros review I said you buy the company, not just the software, for three to five years. That cuts the other way here.

On the things you can actually observe, Discourse looks like a healthy vendor. They are putting out new product on a steady cadence, the AI work especially. They are actively winning new enterprise logos, which is exactly the signal I told you to look for and exactly the signal Khoros is not giving. And the team is technically excellent, and people like working with them.

There is one thing to weigh. Discourse also cut staff in late 2025, which they framed as adjusting their cost structure for long-term sustainability. The difference, and this is my read, is the situation around it. Khoros cut deep after a private-equity acquisition while customers were leaving. Discourse cut from a smaller base and has kept shipping and winning customers since. Same action, very different context. I would weigh the context, not just the headline, and I would ask any vendor, including this one, what their team looks like today.

That is why Discourse is still one of the three platforms, alongside Vanilla and Gainsight, I see Khoros communities migrating to most often. The discussion quality is part of it. The bigger part is that the company behind it is pointed in the right direction, shipping toward where the market is going rather than away from it. For a multi-year bet, that matters as much as any single feature.

So Who is it For?

Buy Discourse if you recognise this:

  • Discussion is the heart of your community, and you want the best discussion experience on the market.
  • Your members are technically minded, developers, gaming, power users, the audiences Discourse was built for.
  • You have, or can get, a little technical resource, and you are willing to actually use the platform’s depth.

Look elsewhere if your community runs on events, if product ideation is the main job, if board-level analytics out of the box are non-negotiable, or if you need a stripped-back, simple experience your members can use without any learning curve. Discourse is a deep forum, not a simplified one. For the right community that is the whole point. For the wrong one it is too much.

Price and Tiers

On commercials, Discourse is more transparent than most of this market, which is refreshing.

They sell hosted plans in clear tiers. Pro and Business are publicly priced, roughly a hundred and five hundred dollars a month respectively, and the full AI feature set is available across them, not locked to the top tier. Above that sits Enterprise, which is custom-priced and where you would land as a large organisation, with the security, scale, and support a big community needs. For that one you talk to sales and get a current quote, and as with every platform in this series, do not take a number off a slide, including mine, because it moves.

There is also a self-hosted option for teams who want to run it themselves, but for an enterprise the realistic conversation is the Enterprise tier. To give you a rough anchor, enterprise and RFP-level engagements have started in the low thousands of dollars a month with a dedicated success manager attached. Treat that as a floor rather than a quote. The point worth taking away is that the lower tiers being openly priced tells you something. You can see what you are getting before you ever speak to a salesperson, which is not true of most of the heavyweights.

The Verdict

Discourse is the best discussion platform in this series, full stop. It scores a 9 on discussions and an 8 on both AI and moderation. If your community is built around conversation and knowledge, especially with a technical audience, it is at or near the top of your shortlist, and the AI readiness is a genuine reason to prefer it right now.

It is not an all-rounder in the way Vanilla or Gainsight are, and it is not trying to be. It is a deep specialist. Events and ideation are its weakest areas, and content, mobile, and analytics sit in the middle of the pack rather than out front. That is the honest picture.

So rank it like this: number one for discussion-led communities, especially technical ones, a strong choice for anyone leaving Khoros, and the wrong tool if events, ideation, or out-of-the-box analytics are the point. If you are leaving Khoros and discussion is your core, this is the first platform I would look at. Just go in committed to using what makes it special.

The full scorecard behind this series is free. Download it, check it against your own shortlist, and tell me in the comments where you think I have got Discourse wrong. And if you want help choosing a platform, or getting off Khoros cleanly, you can find me at feverbee.com.

Subscribe for regular insights

Subscribe for regular insights