01 / 2018
Confidential compute for Web3
The founding team starts from one premise: computation should be private without becoming unverifiable.
stars by this point
0
source
pre-public GitHub signal
PHALA MANIFESTprivate compute · public proof
Dear curious internet wanderer, Welcome to Phala. We build verifiable private compute. AI is moving into agents, model APIs, data jobs, and GPU clusters. The next interface is not a page. It is a runtime making decisions with private context. Those workloads carry prompts, keys, memory, weights, and customer data. Cloud logs, host shells, API traces, and admin panels should not become the trust model. Private execution should not mean trust the cloud account. A result should bring hardware receipts with it. Proof should travel with the answer, not sit in a dashboard only the provider controls. The receipt binds hardware, runtime, image, compose, and output. It should be useful to a developer, an auditor, a user, and another agent. We want private AI to feel normal to build, and difficult to fake. Builders should be able to verify what ran without seeing the secret. That is why Phala exists: public proof for private AI.
Build history
GitHub stars by year.
active year
2018
Phala-Network1,263
Dstack-TEE/dstack480
Calculated total1,743
2018
0
2020
105
2024
1,049
Now
1,743
milestone highlight
research direction
02 / 2020
TEE network goes live
Phala ships decentralized TEE infrastructure and begins turning hardware attestation into a product surface.
stars by this point
105
source
Phala-Network + dstack
03 / 2024
GPU TEE and AI workloads
The work expands from CPU enclaves into private inference, GPU capacity, and applied AI deployments.
stars by this point
1,049
source
Phala-Network + dstack
04 / Now
dstack, Phala Cloud, public proof
The current stack packages confidential runtimes, deployment tooling, and verifiable AI paths for production teams.
stars by this point
1,743
source
Phala-Network + dstack
What the company is built around
A small team building the proof layer private AI needs.
Team offsite
Product planning
Builder session
Conference floor01 / Root of trust
Proof before platform trust.
Attestation, open source runtime paths, and public evidence are the product boundary.
02 / Builder path
Make confidential compute usable.
Docker apps, GPU jobs, model APIs, and agent workloads should deploy without bespoke security plumbing.
03 / Ecosystem
Portable across teams and clouds.
Phala works with AI, data, GPU, and Web3 teams that need verifiable private execution.
Team
Operators, engineers, and ecosystem builders.
A compact group working across research, product, engineering, developer relations, sales, and community.
Join Phala
