Build and run safe AGI on Phala Cloud— a zero-trust confidential compute platform for AI teams. Spin up TEE CPU & GPUs in seconds, keep AI encrypted end-to-end. Get 30 days of free credits on launch.
👋 Hey Product Hunters! I’m Marvin, founder of Phala Cloud.
Why we built this
AGI is coming faster than most of us expected, but it can’t be safe if it runs in an unsafe environment. We’re carving out a “safe zone” in the cloud—confidential by default, open-source from day one—so teams can accelerate toward trustworthy AGI without hand-wringing over who can peek at their models or data.
What’s live today
> GPU TEE for secure LLM training & inference > CPU confidential VMs for any workload > 1-click deployment UI (60 s from idea to running) > Zero-trust key management built into every VM > Automatic security report & proof for each app > API & CLI for full automation > Open source core (transparency > marketing copy) > Free trial—spin it up, break it, tell us what hurts
We’d love to hear
“If your users trust the AI, you unlock the market.” So…
What proof or guarantee would make you trust an AI-powered product? Which part of your stack—or which model—would you move first into a confidential cloud? Any friction you hit during the free trial? Let us know and we’ll fix it in the open.
Drop questions, hot takes, or feature requests below—our team’s here all day. Thanks for helping us push toward safe AGI at cloud scale! 🚀
@marvin_tong Marvin, I always found the "zero-trust" moniker as contradictory. It literally sounds like "I don't trust you" vs "I don't need to trust you"
also markdown doesn't work in this section -- this clearly looks like a copy paste job from an LLM, so don't that not a good look.
Phala Cloud’s zero-trust cloud for secure AGI sounds powerful and timely. Supported it. Congrats on launch – I'm launching my product too, feel free to check it out.
Phala Cloud
👋 Hey Product Hunters! I’m Marvin, founder of Phala Cloud.
Why we built this
AGI is coming faster than most of us expected, but it can’t be safe if it runs in an unsafe environment. We’re carving out a “safe zone” in the cloud—confidential by default, open-source from day one—so teams can accelerate toward trustworthy AGI without hand-wringing over who can peek at their models or data.
What’s live today
> GPU TEE for secure LLM training & inference
> CPU confidential VMs for any workload
> 1-click deployment UI (60 s from idea to running)
> Zero-trust key management built into every VM
> Automatic security report & proof for each app
> API & CLI for full automation
> Open source core (transparency > marketing copy)
> Free trial—spin it up, break it, tell us what hurts
We’d love to hear
“If your users trust the AI, you unlock the market.”
So…
What proof or guarantee would make you trust an AI-powered product?
Which part of your stack—or which model—would you move first into a confidential cloud?
Any friction you hit during the free trial? Let us know and we’ll fix it in the open.
Drop questions, hot takes, or feature requests below—our team’s here all day. Thanks for helping us push toward safe AGI at cloud scale! 🚀
Llanai
@marvin_tong Marvin, I always found the "zero-trust" moniker as contradictory. It literally sounds like "I don't trust you" vs "I don't need to trust you"
also markdown doesn't work in this section -- this clearly looks like a copy paste job from an LLM, so don't that not a good look.
Phala Cloud
@ari_nakos Thanks for the feedback! It's quite helpful 🙏
UI Builder - Mockup tool
Phala Cloud’s zero-trust cloud for secure AGI sounds powerful and timely. Supported it. Congrats on launch – I'm launching my product too, feel free to check it out.
Mukh.1
Phala Cloud looks seriously powerful! We just launched Mukh.1 too - AI agents to handle your everyday ops. Would love your feedback!