Cryptographic proof systems for AI, defense, and enterprise infrastructure. Autonomous action you can trust. AI execution you can prove.
@LagrangeFndn$LA
AI doesn’t need to be a black box.
So we opened ours.
DeepProve is now open source. The fastest zkML in the world.
Fork it. Run it. Break it. Build something weird.
→ lagrange.dev/blog/deepprove…
🔴 AMA: Allora x @lagrangedev
Thursday, June 11 · 3PM EST / 20:00 UTC
Allora's Head of GTM @MichaelZski will sit down with @BrianNovell from Lagrange labs to break down how ZK coprocessors make AI verifiable, and what it takes to prove an Allora inference end to end.
Got
Finally deepprove is open sourced 🎉 the repo is here github.com/Lagrange-Labs/… and you can run your own bench for gemma3 and gpt2 (and llama2 but not optimized). It's quite modular so adding another model is easy ! Reads safetensors, gguf, onnx !
AI doesn’t need to be a black box.
So we opened ours.
DeepProve is now open source. The fastest zkML in the world.
Fork it. Run it. Break it. Build something weird.
→ lagrange.dev/blog/deepprove…
“60× faster zkML” sounds like something marketing would say.
Unfortunately for marketing, the benchmarks are public.
Fork it. Run them. Beat us if you can.
→ lagrange.dev/blog/deepprove…
1/ After 1.5 years of focused work with a dedicated team at @lagrangedev, I am thrilled to announce that our central project, DeepProve, has been accepted at CCS 2026 and is now fully open-source. See paper here eprint.iacr.org/2026/1112 and code here github.com/Lagrange-Labs/….
Thrilled to announce that DeepProve is now open source.
Engineers, researchers and enthusiasts will now have access to the fastest zkML proof system.
The black box is open.
Build with it and feel free to reach out if you have any questions.
AI doesn’t need to be a black box.
So we opened ours.
DeepProve is now open source. The fastest zkML in the world.
Fork it. Run it. Break it. Build something weird.
→ lagrange.dev/blog/deepprove…
Three years ago, we made a bet.
That cryptography could do more than secure rollups.
That it could become part of the trust layer for AI itself.
Today, every line of how we built it is open.
DeepProve is now open source and available to everyone.
Yesterday we open-sourced DeepProve.
Today, the proofs keep running:
12M+ proofs generated 3M+ AI inferences verified 200+ integrations across the Lagrange
network
Open code. Real proofs. Even agents.
→ lagrange.dev/blog/deepprove…
Trusting AI Agents: Mistake or Necessity? 🤔
AI After Dark is back with @lagrangedev and @AIVM_Network!
We're discussing identity, reputation, and what it takes to verify agents as they become part of the economy.
🕑 14:00 UTC
📅 Thursday, June 4th
📌 x.com/i/spaces/1nGeL…