In late September, Turkish entrepreneur Emir Karşıyakalı took to X to show off Claude S40, an app he vibe coded to make his old Nokia 6300 a bit more useable in 2026. In its original form, his app app allowed the 6300 to connect with Claude over a Java and Go backend. Since then, Karşıyakalı has updated the software to make it work with all of the major chatbots, including ChatGPT, Gemini and Grok, and renamed the app Aikon — that’s Nokia spelt backwards — to reflect the greater compatibility.
Outside of a phone running either Nokia’s S40 or S60 operating systems, you’ll need a Go server to hold the API key from Anthropic, Google, OpenAI or SpaceXSI. This part of the equation is essential since it’s the server that communicates with Claude, ChatGPT, Gemini or Grok, and truncates their responses into 8kB blocks your old phone can read. It’s also worth mentioning, since you’re going through each company’s API, you’ll need to set up token billing.
For what is ultimately a gimmick, Karşıyakalı’s app is quite robust and thoughtful. There’s an reading mode, with a dedicated keypad shortcut, to make using a chatbot on the 6300’s 240×320 display a bit more manageable. Provided you a memory card, you can also save responses to the 6300 for offline reading. If you want to send photos to your favorite chatbot for analysis, you can do that too. It’s even possible to send voice messages up to 30 seconds long, with an OpenAI model handling the speech-to-text conversion. Over on the server side, Aikon will hold messages for up to 30 days, and the app supports English, Spanish, Portuguese, French, German, Russian, Indonesian and Turkish. Karşıyakalı has open sourced the app so others can continue building on the app.
Now, before you dig out your old Nokia phone, just know if you live in the US there’s no way you’ll get this app running. That’s because the nation’s last 2G network went offline this summer when T-Mobile shut down its old GSM network in August. If you live outside of North America and want to try Aikon out for yourself, head over to Github and follow the installation instructions.