Codebender Shuts Down was a cloud based IDE for Arduino development. It was made for hackers by a few fellows in Greece. Unfortunately, while they saw some serious success, they were never able to convert it all the way into a viable business.

By November 31st will be completely shut down. They assure users that the site will be in read-only mode for as long as the end of the year, but longer if the traffic justifies it. Codebender made it all the way to 10,000 monthly active users, but hosting and administration overshadowed this success to the tune of 25,000 dollars a month. Not so much as far as businesses go, but without revenue it’s more than enough to shut down a site. Their business plan aimed to tailor their services for specific chip manufacturers and other companies but those deals never came together.

It’s a pity, we were excited to see if Codebender could continue to grow. They were certainly doing some really interesting stuff like remote code upload. As the comments on the site show, many users, especially educators and Chromebook users, loved Codebender — your code isn’t stuck on one computer and where there was a browser there was an IDE.

Two paid services will remain (starting at $10/month) at addresses with different TLDs. But the post does mention that the Codebender project started as Open Source. Their GitHub repo isn’t a clear path for rolling your own, but if you do manage to hack together a working Codebender implementation we’d love to hear about it.

66 thoughts on “Codebender Shuts Down

    1. Always the way.

      I do wonder then if someone might cobble together a service based on something torrent-like. While we’ve seen people using even the blockchain for data storage and archival, we haven’t seen (afaik) services that store their data in (encrypted) pieces on other user’s machines, perhaps even balancing storage usage across the userbase. Here, the ‘cloud’ theoretically has no single point of failure. I would especially love if someone made a library/plugin for this, so a variety of applications and services could rely on the technology, which need low-volume distributed data storage and backup that’s resistant (like bitcoin) to people trying to manipulate it.

      The sketches used in programming an arduino aren’t really that big. It might also make an interesting host for any generated database of cell site sims, which nobody powerful could just go and mess with.

        1. Raises a good point.

          Though the best way to keep CP and other shady stuff out might be instead of encrypting the data, just signing it to ensure integrity. (which actually brings this closer to how bitcoin functions, except swapping transaction data for generalized database entries) Then it’s out there and otherwise the same, except readable by everyone. That wouldn’t make a difference to projects needing a public database, but anybody wanting to distribute the nasty stuff (CP, pirate DVDs, regular porn ect.) wouldn’t be able to hide behind the data being unreadable. It wouldn’t stop them from still encrypting the data, but then it’s easy to find, being encrypted and all. (just a thought: if the data were always stored with the name of the project in some kind of header, their data would be the only data that either didn’t list the project, or couldn’t be decoded (as in unpacked into a readable format) using software from their ‘project’)

          The next question is, why not just use bitcoin? Bitcoin is certainly usable, as has been demonstrated on-and-off, but it doesn’t seem to be optimized for the task of general data storage at all and would require a lot of wrapping in protocols would preform the task of breaking up and translating the data for storage, and retrieving it and putting it back together into a particular project’s reccords.

          1. I think in this case you’ve ended up describing IPFS, or at least something similar. I’m not sure how easy it’d be to make something like an IDE that you could share things on, but I suspect with a little help in the browser (to handle the creation/uploading) you could do it pretty easily.

  1. The challenges of open source and making money.

    Then the whole cloud thing (roll eyes)

    When will people learn. Centralisation (“the cloud”) has some very nice management advantages, everything in one place, ecconomical, easy to control. The reality is it doesn’t work – weather it’s a business or government or manufacturer as soon as you put all your eggs in one basket it doesn’t take much to break the system.

    We need to get off the idea that every thing should be free in that we don’t have to pay for it. Charge a fair price for your service/product/INFORMATION and we all stay in business.

    Theres the challenge of making a system to easily charge and pay for services provided. But one that needs to be implemented.

  2. just to clarify, we’re keeping the two sites up for people who want to keep using chromebooks and they have no alternative, and we’re focusing our efforts on B2B deals with chip manufacturers

    we’re leaving the Maker market cause we can’t justify spending $25K on that anymore. it’s become clear to us that there is no sustainable business model there, or at least we can’t find one. it would make sense for a big company (i.e. Arduino, SparkFun, Autodesk, Make Media) who can justify it based on their business model (they gain from having the community, make money some other way like Professional products or Hardware sales), but it doesn’t for us

    1. So because you didn’t make enough money to sustain your cloud, you are showing big “F**K YOU!” to everyone who used your service for free. You suck at both math and business modeling. Next time don’t offer free, cloud-based service if you can’t afford to keep it running indefinitely…

      1. It’s a business model with precedent (see: Facebook/Twitter); build a service that provides value, figure out how to monetize later. If the monetization doesn’t pan out then the business folds, as its doing here. It’s not like they can continue to hemorrhage money indefinitely; without adequate inflow, eventually funding dries up, and creditors often want their money more than they want a graceful shutdown.

        1. Except Social Media already had a business model :”Convince people your site is best, & sell their personal* data to advertising/marketing companies”. When the sourcecode is out there for free & your target audience is people who would rather spend an afternoon bitbanging or debugging their own code, what are you going to sell? It’s not like RedHat or Ubuntu where you give away the code but sell tech support. Free forums with community driven volunteer techsupport is the goal. They don’t even sell hardware to recoup their investment like Arduino (whichever you choose) or RasberryPi (which also gets loads in public grant money). Even if sites like Twitter built their wings on the way down, no one can honestly think that’s a sound business model.

          *Whether it’s actual personal data, Age, sex, location (market demographics are big money), or just shopping habits (yelp, foursquare) there was already a market for what MySpace, Facebook, & twitter were selling. Being able to streamline 3rd parties into your API (ie; foursquare in your facebook profile) is where the money’s at.

        2. It’s lame excuse of business model based on every kickstarter campaign ever. It’s hope-based model: lets start a product/service and hope someone richer than us buys this stinker before it collapses under its costs. And Facebook is bad example, because they started to sell their data and put ads to get money to run their servers. These idiots didn’t learn from that example, and didn’t learn anything about planning and running business. They hoped some manufacturer will give them money for support for his chips. They didn’t check out first if any manufacturer would be interested, they didn’t plan any contingency plan in case it didn’t worked out, they didn’t plan any sustainable source of income. They suck both at math and business modeling…

          1. Moryc, your words are a prime example of Hackaday’s notorious reputation for haters and hateful comments. Even if your points have merit, why must you be such a hater?

          2. +1 … not for Moryc but rather for Paul Stoffregen’s classy response. Thank you, Paul!

            Moryc: please take your anger elsewhere. We don’t need it on Hackaday.

          3. Sometime you have to try things out to determine whether they are worth it or not – like for instance, I had to read your comment before realizing you are a big hairy scrotum who is a poor excuse for a human. Had I known that ahead of time, I wouldn’t have bothered.

          4. What I said is not less valid or less true because how I said it. I was harsh, I’m sorry. The only thing that I hate is stupidity…
            I’ve seen few businesses that failed because no one planned or modeled them, and everyone assumed best case scenario is their future. A friend of mine was hit pretty hard by that wrong assumption. My brother almost followed with his own crazy idea for a business. Fortunately bank refused him a loan.
            Free service works only when provider sells something on the side. Like ad space, personal data or related products. Everyone knows that (except for Codebender people). You can’t provide quality service for free unless you planned an income source connected to it. Spending 25k USD to provide free service for thousand people without any source of income is just throwing money away. Big companies can do that because they are big, true. But they started small and became big, obtaining a valid source of income. Hope-based business model doesn’t work. That’s why so many start-ups fail.
            Even cloud-based service with associated, expensive product is a bad idea, like that home automation thing that became junk because company went out of business. Why? Because to keep providing service they had to keep selling their product, and sooner or later someone else will start doing the same thing. They can either spend more money on marketing or make their device cheaper. Or do nothing and hope someone will buy their IP before they had to file for bankruptcy.
            Codebender was doomed from the start, by its creators, who didn’t plan a sustainable business. Let’s hope they learned something from that. Many people don’t…

          5. Dear Moryc why do you need say things like these? Every company starts out with an idea, maybe a dream, and tries to make it work. Some are successful some aren’t. Sometimes the market changes. Even the now super powerful Apple was on the verge of failing back in 1997.
            Hackaday really needs to do something with this constant hate from people who contribute ZERO to the community but have a vitriolic opinion about everything.

      2. Probably feeding a troll, but they have to eat as well. So what “free” have you committed to help provide? You should be pleased that this meal I have provided you doesn’t cost you a thing Come tho think of it it didn’t cost me a thing to send it to you; far out love and peace and all that stuff.

      3. Calm down.

        the code is (reportedly) out there on Github. The codebender people are upfront about it and doing the best to ease the transition pain.

        Just because they failed at a viable business model for a free service doesn’t justify insulting them.

        Pick up where they left (they give *you* the means for that) and do it better, instead of whining.

    2. Okay, I have to ask: what exactly is costing $25k per month? That sounds like a lot more than I would expect.

      I’m seriously asking. I’d be interested in taking it over if I can see how to make it cost effective.

        1. That’s about what I figured it’d be. I’m sure they could optimize the amazon costs to get them down with some work (maybe down 2-3k easily), but that’s still kinda a drop in the bucket compared to the rest that just won’t go down that much.

    3. I’m really annoyed at all the negativity thrown your way, Sorry about that you tried a concept a didn’t pan out and you moved on. It isn’t great but it seems like a much smarter move than bankrupting yourself to appease the “I expect everything open source or free crowd.” Well done for at least giving it a go.
      And to all the haters spewing crap go make your own service and thrown $25k of you own money at it, when it fails just keep throwing money at it at least you will beholding to your principles.

  3. Anyone feeling a need to applaud at this defeat, be sure to include a link to the project you worked on that helped people learn to use microcontrollers, even if their computer’s OS was really just a browser that was just an experiment to steal market share from tablets. I mean, just so we can actually assume you’ve a horse in the race and the users can expect to migrate to something better, rather than just thinking of you as a lost YouTube comment section talking head.

    I mean, this helped people a lot more than a comment jerking one’s self off about how cloud computing a buzzword to justify trusting someone else with your computing and Arduino is overcharging and a blasphemy against 555 timers. Here’s hoping the code can be forked in a way to allow Chrome OS users to continue to doing something beyond passively consuming the web.

      1. I read a lot of hackaday comments as a hazard of my work so let me translate.

        “If you are being a turd about this, step up and show us your cool project in a similar vein to this one. This project was really useful for people who owned chromebooks. Those users have lost something they liked and it is not a good thing. It’s important to me that you be a member of a community who participates in and understands that community in a positive way before you arbitrarily weigh in on that community with negative criticism. (non sequitur) Marketing statistics dehumanize us.

        Your comment didn’t help anyone while this project helped lots of people. We all know about the cons of cloud computing and you’re just being a loud meany to bring it up. The arduino is overkill for some projects, but trusting a cloud service with it is not a big deal.

        I hope this software can continue on as an open source project. I feel that more people learning useful skills is a good thing and that tools can help with that.”

  4. I tried Codebender a while back (at least I remember doing so) and I just don’t get WHY anyone wanted this. I think it’s neat but I could not imagine where I’d actually need this sort of service. I think services like 123 Circuits are neat given you can program a simulated Arduino and hook up virtual sensors. That makes it great for beginners.

    Arduino Create is also ‘sort of neat’ but I find myself not using it at all.

    I’ve gotten used to using Visual Studio/Visual Micro and it ends up being great. Large projects are easy to see. I can refactor my code. It naturally works with Git so I can keep versions of projects around.

    Maybe I am just not ‘trying’ hard enough to find a use for them.

    1. I really wanted it to support my 3D printer’s board. For that it would have been a killer app for me, however it was kind of convoluted to get it working and I just moved back to the Arduino IDE. I think too if they had a sort of template support for arduino code. Like you could change the defines for any random library using a nice web interface, but there was just an editor and one with less than I needed. So I signed up, but didn’t use it. Also I don’t like bright white IDEs. Hurts my eyes. That’s just a petty personal thing though.

    2. Sorry to hear of the demise of codebender. Not many people are in a position to offer a free service that costs $25k/month to sustain.

      I’m late to the Arduino party … just moved over from PICs this summer, but already I’m confortable with the Arduino IDE, and the vMicro /VS combination.

      I guess the “cloud” idea is good for any beginners looking for the easiest possible introduction, and this would also be great for educators – no student installs to to struggle with. Here’s the URL, start developing.

          1. The Arduino IDE installation is basically unzip the file into a directory and then execute arduino.exe (at least on Windose). I cannot imagine a more simpler install.

        1. Sometimes students can be using computers where installation of software is forbidden like libraries classrooms etc, So I can see the market I just don’t think it is big enough sadly.

          1. Or maybe it is just too soon. I think codebender could be used a lot in K-12 but not yet. Especially with chromebook support. It’s just recently that more schools are doing 1-1 so I think it is a matter of time where electronics and code will be taught in K-12.

        2. There are a ton of use cases for this. If you’re in a class where installing all the software would waste a lot of time, a cloud based software saves a lot of time. For example we’ve had good feedback on Arduino Create from people who are using windows and the classic IDE is slow while compiling on linux in the cloud is faster. Other use cases are platforms where the toolchain weights several hundred megabytes and might be available only on windows. the list if very long…

          1. Mbanzi, for other platforms it WOULD be great. We teach Robotics on the EV3/NXT using RobotC. RobotC only runs on Windows and I’d love for there to be a cloud version so we could toss Windows out the door for good.

            The Arduino IDE & toolchain is ubiquitous. One can compile on Windows/Mac/Linux. I’ve even seen a version someone made to compile and download from an Android phone/tablet. That’s why in my original post I found an online service not as useful, at least for Arduino.

            Though Autodesk’s 123D circuits (aka has the distinct advantage of adding in components and sensors and allowing them to hook them up to a virtual breadboard. That I find is VERY useful for teaching.

      1. 37% operating after 4 years in information sector, based on your data. If we use this as future indicator, after 8 years only 13% will still be running, after 12 yrs 5%, 16 yrs less than 2%. Looking at the big tech companies this looks like realistic scenario: gone are Atari, Commodore, Sinclair, Acorn, DR, Borland, and so many others. Who survived? Unscrupulous ones like MS and Apple.

  5. Surprised at the anger this announcement generated.
    I applaud the effort and the will to try. If this was a Star Trek universe this would have succeeded. Too bad we’re a planet populated by Ferengi.
    Maybe it could work if it was part of a paid course such as Coursera or one of the others and was used as part of the course.

    1. Codebender was free. The IDE is free. Even using Visual Micro on top of Visual Studio is free.

      Yet you claim we live on a Ferengi planet?? If anything we’d be bargaining on an individual basis with gold pressed latinum to use any of these products.

  6. There is something completely wrong with $25000 per month for 10000 active users per month service. Any $20 per year VPS will easily handle 10000 active users per month.

    If $20 per year VPS can’t handle your 10000 active users per month service, there is something completely wrong with all that crappy web2.0 etc. stuff.

        1. Good find, I knew salaries where the killer but it is amazing how the other “little things” add up. I think the way micro controllers are getting more mainstream this might work in the future. Kids using BBC:Micro or clones in other countries.So perhaps they were just a little a head of their time.

    1. node.js baby, _every single_ static page is painstakingly generated by javascript running on the server from data stored in some SQL database
      this is how simple website can eat few 16 core hundreds of gigabytes “cloud” servers.

  7. As part of a team of archivists who archive web content (Archiveteam, headed by Sketchcow). These guys actually contacted him letting him know they’re going down so we can move in and grab the content nicely before the end.

    This is awesome. It’s a shame to see the site go, but the fact they’ve reached out means that the content will be on at some point, in some usable form.

    1. Before they go, they should leave instructions on how to create your own (local or on-site) codebender-esque site/server. I loved the fact that i could install the drivers on my machine, and then just plug a micro into the computer, and code away, with my stuff saved to the web/cloud. Plus it seemed to work(compile) a LOT faster than the arduino IDE ever did.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.