Future PresentLabs
AI / Engineering & governance
Move fast and make things.
Open-weight engineering models. GPU inference. Agents with explicit authority.
Open-weight engineering RLCD models
Verk
Open-weight engineering intelligence.
From design intent to verifiable outcomes.
Explore our models ↗Verk release in development.
AI & compute governance
Your AI.
Your rules.
Choose where intelligence runs, what it can access and which actions it can take. Give agents the access they need—not the keys to everything.
Keep critical decisions under your control, with explicit permissions and validation before execution. Open foundations you can operate yourself. Managed infrastructure when you want it.
Inference infrastructure
Models need more
than an API key.
One connection to hosted models and your supported AI provider accounts. Choose the model that fits the job—and whether inference runs on your hardware, our infrastructure or through your provider.
- Open weights
- Run supported models on your own hardware or through FPL. Check each model’s license and hardware requirements.
- GPU-backed execution
- Model serving, routing and agent execution are separate concerns. We confirm model, accelerator and deployment compatibility during onboarding.
- Know who you’re paying
- For FPL-billed models and GPU inference, add credit before use; usage draws down that balance. When you connect your own provider account, its pricing or plan limits apply. Any FPL hosting or service charges are separate.
FPL Workbench · In development
Engineering tools.
One workspace.
Transmog works on geometry-to-CAD workflows. Legion-of-BOM brings component sourcing and bills of materials into the engineering process. Workbench is the shared interface we’re building around those tools.
Transmog Legion-of-BOM
Explore engineering →
TTS-JARVIS · Open weights
Give your AI
a voice.
A little more character. A little less robot.
Quick listen selected · Press play
Recorded examples. Pick a sample, then press play. Custom text is not available in this demo yet.
VoxCPM2 LoRA adapter · Apache 2.0
Decision Catalog Protocol · Draft
Give AI choices.
Keep authority.
A live catalog of allowed actions. AI selects; your system validates and executes.
Read the DCP protocol ↗
Research & business intelligence
See the signals.
Keep the evidence.
Connect research and operational data to decisions, with evidence you can trace.
Explore our intelligence work →Research Wing
Exact sources. Cited evidence. Reproducible comparisons.
Panopticon
Data and business intelligence, grounded in context and evidence.