Amateur Repeater Directory

[K9KLC]

I will also share it with fellow club members with the caveat of inaccurate results due to errant, or nonexistent data points on which the initial website was developed. “Bad data in, bad data out”. Hopefully operators will start making improved/accurate edits for their respective areas.

And this is how we get things more accurate. It's always been "Garbage in = Garbage out".

 

[N4DES]

I'll be honest within 15 miles of me, there's probably 50% or better wrong. I didn't do a count but when there are 10 repeaters or so listed and 7 of those are not on the air and have not been for quite some time that's a pretty good rate of "incorrect". I'll submit more later but a lot of what's listed is NOT on the air now. Also one is listed in the wrong location. (probably the widest area UHF repeater near me). More corrections to follow.

Again, I'm NOT complaining or anything. These things can only be as good as user input is.

73-Greg

The initial data that was used for the repeaters around me is at least 10 years old.

 

[AK9R]

I'll be honest within 15 miles of me, there's probably 50% or better wrong.

There was a time when over 50% of the listings for my county on a well-known online amateur radio repeater directory were incorrect. Now, the inaccuracy "rating" is about 33%.

 

[KJ5GZR]

I have a repeater going up soon so this will be very helpful. I've been using Repeater Book to find repeaters, but this may be better for me in the long run. Awesome project! I'll spread it in the local clubs.

 

[K9KLC]

There was a time when over 50% of the listings for my county on a well-known online amateur radio repeater directory were incorrect. Now, the inaccuracy "rating" is about 33%.

At least it's getting better not worse.

 

[CoastalDude]

Okay, We are approving peoples changes now, lot's of activity, over 200 users signed up in the last 24 hours.

What state are you looking at K9KLC and I'll let you know why these are incorrect. Some of the states we gathered a year ago, but now I use AI and it's FAR faster and WAY more accurate.

Again, let me know what state you are looking at and I'll see what I can do

 

[CoastalDude]

I have a repeater going up soon so this will be very helpful. I've been using Repeater Book to find repeaters, but this may be better for me in the long run. Awesome project! I'll spread it in the local clubs.

Thank you, and you can make edits, we need to get the data up-to-snuff, it will requires the communities help however. You can edit repeaters by selecting the repeater, clicking the blue "Properties" window, and you'll see the little icons for "Add" and "Edit" in the top right corner of the properties window. Once saved, they get reviewed and once approved go live, shouldn't really ever take more than a few hours for that to happen

 

[K9KLC]

What state are you looking at K9KLC and I'll let you know why these are incorrect. Some of the states we gathered a year ago, but now I use AI and it's FAR faster and WAY more accurate.

Again, let me know what state you are looking at and I'll see what I can

Illinois in Madison county. Again the callsign "AA9MZ" has NOT had a single repeater on the air since I've been licensed. (1995) there's a story behind that stuff but I'll not go into that here. The callsign KG9OV (224.7) you do show now as off the air, I talked to him and he'd rather it not be on there at this time opposed to it showing "off the air". I'll submit more thru the site.

but now I use AI and it's FAR faster and WAY more accurate.

But is it? If it's not any better than some of the stuff I've seen it come up with, heaven help us. At any rate, AI is nothing more than taking data compiling it and spitting out an answer. The data originally came from somewhere else, most likely humans were involved. People may know things AI doesn't. (citing above).

 

[kc2asb]

But is it? If it's not any better than some of the stuff I've seen it come up with, heaven help us. At any rate, AI is nothing more than taking data compiling it and spitting out an answer. The data originally came from somewhere else, most likely humans were involved. People may know things AI doesn't. (citing above).

I think this Repeater Directory is a fantastic idea. AI, well, if we just start accepting what AI spits out as gospel, we're lost.

 

[vagrant]

If the AI data results are accepted over user input, this thing is dead. It was fun the few days it lasted.

 

[CoastalDude]

If the AI data results are accepted over user input, this thing is dead. It was fun the few days it lasted.

I use AI all day on this, writing the code for the whole site, I get at least 10 times more done in a day than I would otherwise. Tonight I added a search button that floats on the map in the upper right corner, you can type in your repeaters callsign and it will take you right there, if you have multiple it will take you from one to the next and the next and so on, took an hour, that would have taken me all day if not longer. Repeater owners will love this feature, I started it about 4 hours ago, it's in production now.

When I say AI input, I mean I download the CSV or HTML and feed the data to AI, it knows my SQL table format, it can parse all kinds of antiquated formats now (because ham radio sadly is 1990 when it comes to tech). I can ask it to do things, walk away, come back, plug it in, I test it, find and fix bugs, see how things work, and come up with better ways if need be, that I will also ask AI to write.

The Fresnel zone stuff was mind boggling, I know a sh*t ton about that now, knew nothing before I started this project, and it took a lot to get it right, however, AI was key in getting that done, but in the end, I had to steer AI to the solution, it couldn't get it right without my input in the end.

But yes, once I initially populate a state from that point forward it's basically user maintained, we have people already adding new repeaters and and a lot of edits are coming through. They key to AI, and being good at it (they call it vibe coding) is knowing how to tell it what you want to do, it takes time to figure it out, sometimes AI doesn't get it, a lot of times actually.

The one thing I've learned is the human mind is still far far superior to AI, AI is quick, incredibly quick, knows more about Ham Radio then everyone on this board together, don't think so? Go to chatgtp.com and ask it a question. Yet, the code we've come up with together is far superior because I was driving, some of the logic it comes up with a way way overcomplicated, it won't stop and change the architecture when you're trying to solve a problem, it will keep tweaking until your code until it's incomprehensible, that's when you, the human, needs to step up and say say, no, let's do it this way. Have I cussed at it? Yes.

For instance, I can post an image of a dialog box I want created, ask it to create the dialog HTML in Bootstrap, create an ES6 Javascript class that will contain the code to support the dialog, then to submit that data to a C# Rest API, ask it to actually create the REST API controller, the service and the Dapper SQL to store it in the database. Thus, from creating the UI to getting it into the database, the entire thing, AI can do that in probably 2 minutes, not joking.

But you have to still be a very experienced programmer to do this right.

 

[CoastalDude]

Illinois in Madison county. Again the callsign "AA9MZ" has NOT had a single repeater on the air since I've been licensed. (1995) there's a story behind that stuff but I'll not go into that here. The callsign KG9OV (224.7) you do show now as off the air, I talked to him and he'd rather it not be on there at this time opposed to it showing "off the air". I'll submit more thru the site.


But is it? If it's not any better than some of the stuff I've seen it come up with, heaven help us. At any rate, AI is nothing more than taking data compiling it and spitting out an answer. The data originally came from somewhere else, most likely humans were involved. People may know things AI doesn't. (citing above).

Okay, we can remove repeaters if need be ...

 

[K9KLC]

If the AI data results are accepted over user input, this thing is dead. It was fun the few days it lasted.

Agreed, this is first and foremost. If. my word isn't good enough to make changes over years old data, and the Admins not in the area to check, ya, it's dead. What good is user input if AI takes precedence.

 

[K9KLC]

I use AI all day on this, writing the code for the whole site, I get at least 10 times more done in a day than I would otherwise. Tonight I added a search button that floats on the map in the upper right corner, you can type in your repeaters callsign and it will take you right there, if you have multiple it will take you from one to the next and the next and so on, took an hour, that would have taken me all day if not longer. Repeater owners will love this feature, I started it about 4 hours ago, it's in production now.

When I say AI input, I mean I download the CSV or HTML and feed the data to AI, it knows my SQL table format, it can parse all kinds of antiquated formats now (because ham radio sadly is 1990 when it comes to tech). I can ask it to do things, walk away, come back, plug it in, I test it, find and fix bugs, see how things work, and come up with better ways if need be, that I will also ask AI to write.

The Fresnel zone stuff was mind boggling, I know a sh*t ton about that now, knew nothing before I started this project, and it took a lot to get it right, however, AI was key in getting that done, but in the end, I had to steer AI to the solution, it couldn't get it right without my input in the end.

But yes, once I initially populate a state from that point forward it's basically user maintained, we have people already adding new repeaters and and a lot of edits are coming through. They key to AI, and being good at it (they call it vibe coding) is knowing how to tell it what you want to do, it takes time to figure it out, sometimes AI doesn't get it, a lot of times actually.

The one thing I've learned is the human mind is still far far superior to AI, AI is quick, incredibly quick, knows more about Ham Radio then everyone on this board together, don't think so? Go to chatgtp.com and ask it a question. Yet, the code we've come up with together is far superior because I was driving, some of the logic it comes up with a way way overcomplicated, it won't stop and change the architecture when you're trying to solve a problem, it will keep tweaking until your code until it's incomprehensible, that's when you, the human, needs to step up and say say, no, let's do it this way. Have I cussed at it? Yes.

For instance, I can post an image of a dialog box I want created, ask it to create the dialog HTML in Bootstrap, create an ES6 Javascript class that will contain the code to support the dialog, then to submit that data to a C# Rest API, ask it to actually create the REST API controller, the service and the Dapper SQL to store it in the database. Thus, from creating the UI to getting it into the database, the entire thing, AI can do that in probably 2 minutes, not joking.

But you have to still be a very experienced programmer to do this right.

Thanks for the explanation, much appreciated. Now back to normal user input.

 

[K4EET]

@CoastalDude, I’ll second @K9KLC on the thanks for the explanation. It really is much appreciated. AI is still new and a little scary at times. In the end, the success or failure of the directory will be dependent on your skill watching over the code as well as the database where AI is involved. It appears that you have a good grip on that. Your skill as the coding expert and database manager is paramount. I believe that you are meeting those qualifications not only in these two areas but also in your vision of what will make a repeater directory more valuable to the end user. I pray that the end users will help make the data more accurate than it has ever been before. Then we will all have a valuable resource. Maybe AI could even “ping” the repeater owners asking for their help with coming up with a more accurate database. You are building a great platform for that database. I look forward to seeing your efforts succeed. In any event, press on my friend and Godspeed ahead! 73, Dave K4EET

 

[AK9R]

Okay, we can remove repeaters if need be ...

This is a double-edged sword which must be handled carefully.

1. There are times when a repeater operator absolutely does not want their repeater listed in public databases. Reputable repeater coordinators honor these requests by flagging the coordination with a "do not publish" designation. But, in crowd-sourced repeater databases, such as yours, if a user provides information about a repeater that should be kept private, what's stopping the repeater from being listed in the database?

OTOH

2. Are controls in place to keep a user from asking for a repeater to be deleted in a malicious action against the repeater operator?

 

[nd5y]

1. There are times when a repeater operator absolutely does not want their repeater listed in public databases.

Sometimes they don't mind the frequencies being published but don't want the exact location of the repeater disclosed.

Trying to make coverage maps without knowing the exact location, height above ground and other system parameters that the average user/crowdsource supplier probably doesn't even know would be a wast of time.

 

[K9KLC]

This is a double-edged sword which must be handled carefully.

1. There are times when a repeater operator absolutely does not want their repeater listed in public databases. Reputable repeater coordinators honor these requests by flagging the coordination with a "do not publish" designation. But, in crowd-sourced repeater databases, such as yours, if a user provides information about a repeater that should be kept private, what's stopping the repeater from being listed in the database?

OTOH

2. Are controls in place to keep a user from asking for a repeater to be deleted in a malicious action against the repeater operator?

I 100% agree with this. Had I not talked to the actually owner ( a close personal friend of mine) I'd never ask for something to be removed. He falls in category #1, he simply said if you can get it off there do it. He said there's nothing worse than still having it listed and shown as "off air" when he has absolutely no desire to have it listed period.

As far as #2 Yep that can happen I guess in this day and age. I forget sometimes people are people and due to the era I grew up in, and the way I was raised I still have problems understanding people like this. I guess we've "come a long way" eh? I still remember waiting for the newest repeater directory to come out and was excited about that. Just like guys that coordinated a ton of frequencies "way back when" and have never had a repeater on the air. They did it because they could not because it was the right thing to do.

There is never going to be a 100% accurate repeater data base. There are a few lower profile 33cm repeaters in my area that are not coordinated yet (like 33cm is so crowded anyway) and I honestly don't know if they ever will be coordinated or made public. Do you post something like that?

Also on to the subject of closed repeaters. I know of several and those guys don't want the word out in any public medium yet, some how they show up and are listed as "closed" and at least one I know of the guy hates it's even out there.

This is definitely a slippery slope, and I, for one, don't have an answer how to handle all of the possible variables that could come into play here.

 

[CoastalDude]

It appears that RepeaterBook is now moving toward selling access to repeater information—data that should remain freely available to the amateur radio community. Many state frequency coordinators have already stopped publishing their databases publicly and instead provide that information privately to RepeaterBook. They already make a profit via ads on the site for data, much of it provided by us, and the frequency coordinators. Was not this data meant to help our Hobby, or was it a for profit venture?

In 2025, the ARRL officially designated RepeaterBook as its primary source for repeater data. Both organizations are private entities, not public agencies, and this shift has effectively placed community-sourced repeater information behind commercial and closed platforms.

This is exactly why we’re building the Amateur Repeater Directory — a transparent, community-driven, and public-domain alternative. Our repeater data is exported regularly and published openly, with no licensing restrictions. It’s truly free as in free beer — anyone can use it, share it, or even build competing applications.

But taking information from state frequency coordinators, claiming ownership over it, and then selling it back to the very community that created it — that’s not how amateur radio is meant to work.

This is exactly why we are building the Amateur Repeater Directory, it's interesting that they are now announcing many of the features we already have.

 

[k6cpo]

Please elaborate on exactly how you derived that 25%. Your calculations and data sources will be helpful.

WAG based on personal observations...