Looked in this forum and on GitHub Discourse for an official Discourse Model Context Protocol (MCP) and did not find anything.
While there are unofficial Discourse MCP (ref), would much prefer an official one. Yes I am considering creating my own if needed.
Update:
DeepLearning.AI just released a free online class taught by Elie Schoppik from Anthropic
We don’t have an official MCP server or client. We have been “waiting and seeing” for now to see how adoption of the protocol emerges.
Would love to have an MCP for Discourse API Docs and Dev Docs. That would help a lot since we have been tweaking some Theme Components and we want to make sure that the code is always compatible to our and future Discourse versions and follow best practice.
Sometimes it is the ideas that you do not think of that are some of the best. 
Thanks for posting.
The “note” seems to be empty when expanded…
It isn’t necessarily an industry standard yet (like JSON, REST, etc) therefore early adopters (organizations/companies included) may end up paying the price of something shinier and better coming along after this original MCP. With that being said, it’s understandable that Discourse may want to wait this one out a bit
For a quick and dirty test I started using GitMCP: https://gitmcp.io/
If anyone else wants to give it a shot, check it out:
My current thinking roadmap wise is is along this… still needs lots of refinement but it helps paint a picture of how I think about this.
Allow admin to select a collection of tools (either custom or built-in) and publish to a particular endpoint. /discourse-ai/mcp/NAME
Use Bearer for authentication - in particular a user API key
Key can be scoped to a particular MCP service
Milestone
Allow admin to consume an MCP service, which adds a list of tools to the available tool list that can then be consumed by personas.
Use Bearer for authentication
Milestone
Milestone
This is a more complex auth mechanism especially registration
Milestone
milestone
leaving this here as a possible way of modularizing it , glad to see a conversation on this topic.
A security issue one should be advised of/know/understand related to MCPs.
Stop Running Your MCP Tools via npx/uvx Right Now by Bob Dickinson
The gist of the issue is that by using npx or uvx you are essentially telling your tool-using platform to go get whatever code is registered with the package name provided and run it (the “x” is for “execute”), updating whenever the package changes. And the code that is then run on your machine has root access — it can see your entire machine, environment variables, the file system, it can open ports to listen or to exfiltrate data. This code, which just got pulled down fresh from the registry, can do ANYTHING it wants.
For example in learning to use MCPs grabbed the time MCP from
The install instructions note one possible option
e.g.
uvx actually downloaded all of the code in milliseconds into a temporary directory and ran it as needed. When the MCP client was done (closed in the case of Claude Desktop), the temporary files were deleted.
Yes you should get a warning before the code is executed as in the case with Claude Desktop
but how many users really know how to check and just click an allow option.
Announcements such as these are common.
Destructive malware available in NPM repo went unnoticed for 2 years
Creating MCP Servers via vibe coding is relatively easy, getting the security correct is not.
@jrgong FYI you can do this today with the context7 mcp which is great. It looks like someone added the Discourse docs too, though I don’t know if it’s API, admin, or both. Either way, you can add what’s missing too.
How do I skip to V10 for full control—messaage bus and all. Plug me in 
Whatever they added is an absolute disaster, Context7 - Up-to-date documentation for LLMs and AI code editors - I read through a few pages of it and it is useless.
Hi everyone,
Just jumping in to say that we currently operate as an MCP client via ravenala.ai, and we’ve seen some of our users start to integrate Discourse through Zapier-based workflows.
We’re especially excited by the potential of a native MCP server, and would love to know if there’s a rough ETA for V0 or any experimental endpoint we might start testing with. Even a simple /discourse-ai/mcp/NAME setup with basic tools (search/read topics, fetch posts) would open up some great async automation use cases.
Happy to share specific feedback and lightweight ideas if helpful—looking forward to seeing how this evolves!
Contrary to what I originally thought would happen, it seems that MCP is being widely adopted and is becoming more mature. May I inquire as to how you guys are now approaching this?
it is 100% on our roadmap now, but no timeline to when we plan to release here
I started working on an mcp here:
Once I make a bunch more progress we can publish it and make it easy to use.
A nice way to show support for Discourse MCP that Sam is working on is to give the project a star on the GitHub repository.
Steps
I am curious, what do y’all expect the MCP to do?
I have the obvious ones of search/read, what are some other specific use cases you are hoping we nail here?
I just added filter which is quite important:
The timing is splendid. I’m working on creating a sort of open source cognitive system to interact with Discourse instances which runs as a daemon with multiple modules. So being able to create topics as well would be dandy although tricky I’m sure. Also, running data explorer queries would allow for enriching context/data quite a lot.
I started building this system using Discourse’s REST API but now I might just wait until the MCP is ready 
then use the API for very niche tasks.
As you noted you have search and the ability to read specific topics, posts, users and tags which covers most of what is needed.
Chat with forum information
One that is of great value is similar to the idea of using AI to read a PDF then chat with the PDF. Instead of the content coming from the PDF, the content is from the Discourse forum.
Now while there is Discourse AI, wrapping some of the read-only functionality with a common interface, e.g. MCP, and being able to access within a common MCP host, e.g. Claude, increases the usefulness.
Additionally with this ability, one could in the MCP host, e.g. Anthropic Claude, add other agents or sources of information that could be used with the information from the Discourse MCP.
Morning brief
On most days I visit over 20 different places for information, a few of them being Discourse forums and looking for specific information, or something of value that I was not specifically seeking. What would be nice is to get all of this automated and a Discourse MCP would be of value; yes I know not all Discourse sites would be required to allow access via an MCP.
MCP Discourse forum discriminator (AKA Relevance score)
As a programmer am a member of several Discourse forums related to software. If all of them have MCP servers and all of them are activated for use with an MCP host, and I ask a not so specific question about coding, e.g. giving a coding fragment, there needs to be a way for the MCP server to give feedback if this question is a good fit for that forum before requesting a more expensive request.
Yes this is an odd question, but we know that in reality this is a possible real world scenario.
Do not expect a solution right away but something to ponder.
Frequently asked question that are unanswered (AKA Lacuna)
Since the AI can quickly scan and summarize all topics into points in the vector space, it is possible for it to identify frequently asked questions that are not being answered. This is a common occurrence when something new is released such as with software updates, or even more real time when servers go down.
Additionally one could use the Discourse MCP to help a user identify commonalities in the information that could take minutes to hours to do so by hand.
From trying to help with server down issues this would be of great value, e.g. a server is working for me but not in another country, or it is working for me with Windows, but not with Apple phones, or there is a pattern of related problems but because users do not use the correct terminology it goes unanswered, e.g. CFG should be used but is not.
For me personally, currently not looking for the ability to write or modify.
