Lesson from TRAE SOLO train wreck: how not to launch a product.

When Perplexity introduced Comet with a waitlist, I thought that was a bad business decision in this day and age where everyone is racing to be the first to market. But TRAE taught everyone that there’s always worse way to do things.

For context, TRAE is another VSCode clone with AI sprinkled on top, similar to Cursor and Windsurf, but they were back by ByteDance (the company behind TikTok) and that got everyone hyped. And for a while they gained a certain amount of cult following. Nothing like the kind of following Cursor or Windsurf have of course, they were quite late to market and initially they were only available on Mac. I used to say that in a world where Figma already gained absolute dominance, TRAE wanted to be Sketch. But they eventually released a Windows version and things started to pick up from there. They had a pretty generous Claude Sonnet rate limit for the free tier and ‘free’ always attract people. The rate limit is pretty hard though just like any other service with a free tier, couple with the fact that they still don’t have a working build for Linux, the platform with the largest amount of developer, means that TRAE remains a novelty of an editor: something people crawl back to at the end of the day when they hit the rate limit in Cursor, Windsurf or Gemini CLI.

But TRAE isn’t just a mere VSCode clone, it’s a clone backed by ByteDance, and that means they have to capital to push it forward, should they wanted to. And pushed forward they did. TREA recently launched version 2.0, marking a complete rebranding, moving away from the color Orange to a more technomatrix style Green color, and that’s not all. They’re also promising a whole new mode called SOLO, supposedly a complete all-in-one agentic mode that can do a lot more than just correcting your vibecode slops, it can also use the terminal, write documentations, use browser etc. And that got their followers losing their damn minds.

Now if you’re like me, you wouldn’t think this is something worth losing your mind over. After all, Cline was already doing it with their Plan mode, Roo Code was already doing it with their Orchestrator mode, even Amazon’s new VSCode clone Kiro is already out the box. TRAE is late to market, again, BUT, quite a lot of their followers don’t know that (remember, their business model attracts mostly free users, lesser developers, vibecoders, especially in China where access to international tools and information is scarce). So when they announced it, almost every single person in their subreddit lost their mind, not knowing of better alternatives.

While SOLO was available as part of the 2.0 launch, and it does come with the app, but it was purposely locked behind an activation code, which you can only obtain by being a paid user and follow them on social media (Twitter/Discord/Reddit). They promised to send out codes to followers after a certain period time. The whole point is to increase their social media following, and it was a huge success. However what followed next was nothing short of a train wreck.

Guess how many SOLO codes TRAE decided to generously offer to the community after hyping up this feature to thousands of developers.

If your guess is 10, you’re correct.

Yep you read that right, 10, TEN, T-E-N, the number between 9 and 11. That 10.

That’s how many codes they decided to release to the public after a week of teasing it in their app. They also decided with this many code, the best way to drop them to the public is by letting one lone community manager, Amber, copy and pasting them manually into Twitter/Reddit/Discord in that order, one-by-one. And as expected, hilarity ensued.

Most of the codes were claimed in milliseconds after posting. They were claimed before Amber has a chance to even press Ctrl V in the next app. None of the people who was excitedly waiting on Discord was able to even get a whiff. And as expected, this turned hype into rage. People flew on social media to dunk on TRAE.

TRAE might’ve gained a few thousands follower on various social media outlets, but they effectively demolished their entire social trust in the process. Was it worth it? Only time will tell.

Case in point: waitlist might be annoying and appears to be pretentious but sometimes it is the necessary evil for rolling out a product in a controlled manner.

Update:
TRAE promised to release an additional 1000 codes and opening up a waitlist afterward.