Skip to content
OpenJev: FREE OpenSource Alternative to TypeSafe Jev! artwork

OpenJev: FREE OpenSource Alternative to TypeSafe Jev!

AI News Today | Julian Goldie Podcast

September 22, 2026

OpenJev: FREE OpenSource Alternative to TypeSafe Jev!

Speakers Julian Goldie

TopicsMarketingBusiness

Julian Goldie (0:00)

Today, I want to show you how to use a free alternative to Jev AI called OpenJev. And I'm going to show you how we actually built out Jevland, which is a self-generating system we can use to play as a game. And then every 10 seconds, it actually regenerates itself based on our progress inside the level, which is pretty cool. Now, you might be wondering, how do we do this? So essentially, OpenJev is an open source alternative to Jev AI.

Once you set it up, as you can see right here, pretty cool.

They've actually got a full tutorial walking you through it as well. And once you've done that, you can actually get a GROK API. So you can get an API from GROK. That's G-R-O-Q.

I'll show you how to do that in a second. From there, it's going to have a setup that runs locally, as you can see right here. And you can switch the models and change it up and switch between parallel and that sort of thing. Also has a bunch of questions and a bunch of pre-built automations, that you see, for example, support routing, content moderation, that sort of thing. But we actually built out JevLite using this system, just for a cheeky bit of banter. So how can you use this? How can you set it up? So basically, the way that I did it is we got the GitHub URL, like you see right here, and then we went into Claude, and I just said, can you test this out and build some cool stuff with it? Gave it some examples of what I built recently with Jev, along with the documentation. I actually set that up for me, as you can see right here. So it runs an OpenJev Playground running locally. And now here's where the magic happens. So we actually gave it a GROK API. Now you might be wondering, how do you get a GROK API and how does that work?

So GROK is like a super fast way to use AI, as you can see right here. It's a lot faster than, say, using something like Fable 5.1 or something like that. So you can use it as the free API inside Jev AI, which means that if the project is open source and you've got a free API that you can plug into it, well, the whole thing is free. You can see how many tokens we've used today using the system. So you can get a free API key. Just go to API keys once you've logged into GROK, create an API key like so. Once you've done that, you can go back to Claude and give that access. And you can see the once we plugged in GROK, I did try originally with OpenRouter, which was a mistake, would not recommend that.

Just use GROK instead. And then this makes it a lot faster, right? So we actually took it from like 850 milliseconds. Sorry, we took it from 6 to 13 seconds on free OpenRouter models, so I don't recommend his and those. Whereas with GROK, we cut the decision time to 0.6 seconds, which is pretty fast, and is still free when you use that. Then it builds out the game step by step. It tests it out, it hosts it locally like you see right here, and that was just kind of a cool test run. Now you might be wondering, okay, how does it work on the backend? How does this testing work? So basically, with this system, there's no pre-made level, right? So when you're playing this, for example, there's no pre-made level, just about every 10 seconds, one OpenJev call reads how you're playing, and then from there it decides the next four sections of six types, the difficulty and the world theme as you go along. The code then builds those islands based on as you go along, and the HUD shows a queue and the last decision's model and latency, plus a banner on each section says whether it came from Jev. It also has a local fallback if the call is late as well, which is super useful. Now, you might be wondering, in reality, like, how can you actually use this for useful stuff? So you could use this, for example, for like a voice controlled computer browser, we've actually already tested that out. You can run it with a system like this, where you can basically make decisions quickly as you can see. And you can even use this with our camera, right? So if we open this up, you can see it pulls me in here. But I'm actually wearing a white t-shirt now, whereas inside the Jev AI app, it's got a blue cardigan, right? So we can switch between the looks, we can add sunglasses, and it's quite easy and simple to set that up using OpenJev. You could use this, for example, for like switching between models. So let's say you wanted your AI to automatically change the model depending on what you're doing. So like it could use a local model or a free model with Jev AI to save tokens, and then if it's something complex, then the prompt could go to something like April 5.1.

1 more minutes of transcript below

Thousands of transcripts fetched by people building searchable podcast archives

Fetch the whole transcript

The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000791174070