We are excited to share that ASTPP Enterprise Edition v7 Beta is now publicly available.
Over the past year, we’ve been working on modernizing the platform - improving the UI, boosting performance, strengthening security, and introducing new tools to help telecom service providers scale better.
We know many of you have been waiting for an update on the community version.
A refreshed Community Edition v7 is also under development and planned for release in March 2026.
We will soon publish the full roadmap along with screenshots and progress updates.
Thank you to all community members, contributors, customers, and partners for staying connected with ASTPP. Your support and patience mean a lot, and we are excited to bring these new updates to life.
Yes, we are targeting March for Community Edition v7.
Enterprise v7 is stable and in production. The CE build is built on the same core and the team is actively working on getting it release-ready. We are doing everything we can to hit this month.
We will keep this thread updated with progress as we go. And if there is any change in timeline, we will post here before the month ends so you always know where things stand. No surprises.
We appreciate the continued interest and questions around Community Edition v7. Here is where things stand.
The team is currently fully focused on two things: migrating existing Enterprise customers to v7, and building the next generation architecture for v8. Both are consuming significant capacity and CE v7 as previously announced has been deprioritised as a result.
More importantly, while working on v7 we made a bigger architectural decision that directly changes what Community Edition will be. We are rebuilding the platform from the ground up for v8 with a fully API-driven architecture.
In this new design, the core of the platform will be open source and will form Community Edition. Enterprise Edition will be built on top of that same core with advanced modules. This means CE and EE will share the same codebase going forward, which is a much stronger foundation than maintaining two separate codebases.
Rather than ship a CE v7 that sits on the old foundation, we are building a CE that starts modern from day one.
We will share the roadmap and architecture details as soon as they are ready. Thank you for your patience and for staying with us.
Doesn’t sound like it to me. It’s odd they are still trying to use the ‘lack of resources’ excuse when you can probably do it in less than a day now with AI agents.
In fact, it wouldn’t be that hard to build a more modern alternative using AI agents now a days. That is probably the future of all software development going forward imo. Everyone just getting Agents to build them custom versions of software according to their specifications.
UPDATE: I had a few spare minutes so I got an agent to read the entire ASTPP code base, create a detailed migration plan, and estimate token usage.
I have tried many models and the best one is Claude 4.7 for sure. You could probably do it on their $20 a monthly plan. Though there is a weekly restriction on how much you can use it.
ChatGPT does not do a very good job at complicated coding. It does well for simple things like Arduino, and other simple programs. I tried the OpenAI Codex and and it does not do well either.
AI models are becoming commoditized. No need to use the expensive ones anymore imo. I use DeepSeek for most of my high token usage grunt work, then I get models like Claude or Gemini to check the work. You can even run models locally on your own hardware that are optimized for coding and supposedly do a decent job.
The people around here need to stop using the “lack of resources” excuse for not doing anything with community edittion. With AI agents it’s almost trivial now.
I tried qwen3-coder-next:q4_K_M locally, and it does not do a very good job, though it was for C++. I have not tested it for other languages.
I only have a 3060 so it exceeds available vRAM, thus a bit slow. I am considering getting one of the AMD Radeon AI PRO 9700 which comes with 32GB of vRAM. It is only $1299. I do wonder if it would be cheaper to use one of those cloud companies where they charge you a fee based on actual usage rather than just having a image sitting idle while still being charged like a tyical VPS.
The agents seem to work best with frameworks, where there is a well defined and documented set of components to work with. The only compiled language I have used with agents is Go, and I was quite happy with the results.
I noticed that Qwen was compacting quite often, so it seems like the context window is small. DeepSeek v4 doesn’t seem to do that as much although that might have more to do with the IDE agent I am using than the underlying model. My understanding is that the more it compacts the more details of the conversation that are lost, so that could result in more mistakes.
When dealing with them, you have to use a file so it does not forget important things. It all depends on the agent/model you are using. For example MEMORY.md (or CLAUDE.md / AGENTS.md)
Just build your own. I would if I had a need for it, but I don’t at the moment. I already built my own alternative to FreePBX. Only took me about a week. I might try build my own alternative to FusionPBX next.
What’s new in ASTPP Enterprise v7 Beta
You can see full details and screenshots here:
Update on ASTPP Community Edition v7