Dust is one of the stronger enterprise AI-agent platforms — and, like osFoundry, it’s model-agnostic, which makes this a comparison of approach more than philosophy. Here’s a factual look for a US business, with sources cited. (dgm implements osFoundry, a separate company’s platform — we are not osFoundry.)
At a glance
| osFoundry | Dust | |
|---|---|---|
| Core focus | Orchestration: agents, automations, apps | Multiplayer AI-agent workspace |
| Models | Bring your own, any provider | Model-agnostic (OpenAI/Anthropic/Gemini/Mistral) |
| Security | Confirmed in assessment | SOC 2 II, GDPR, HIPAA support; EU/US hosting |
| Pricing | Via dgm: $399 / $3,999/mo | Pro ~€29/user/mo; Enterprise custom |
| SaaS consolidation | Designed to consolidate | Connects to tools; agent workspace |
What Dust is
Dust is an enterprise “multiplayer” AI-agent workspace: you build agents, connect them to your company’s tools and knowledge, and collaborate with them in workflows — it bills itself as an “operating system for AI agents.” Founded by ex-Stripe and ex-OpenAI engineers, it raised a 2026 Series B and reports adoption across thousands of organizations. Its center of gravity is a workspace where teams build and run agents together.
osFoundry overlaps here — it also runs agents — but its emphasis is broader orchestration plus the explicit goal of consolidating overlapping SaaS, delivered as an implemented system by dgm rather than a self-serve workspace.
Models
Both are model-agnostic, which is a point of genuine similarity. Dust supports OpenAI, Anthropic, Google Gemini, and Mistral, choosing per task. osFoundry is likewise model-agnostic, applied at the orchestration layer across agents and automations. So neither locks you into one lab — the difference is what surrounds the models (a collaborative workspace vs an orchestration-and-consolidation layer), not model choice itself.
Security and data
Dust publishes a solid posture: AES-256 at rest, TLS in transit, SOC 2 Type II, GDPR, and HIPAA support, with EU or US hosting options, zero data retention, and a commitment not to train models on customer data. For a US business, that’s a reassuring baseline. With osFoundry, dgm confirms the equivalent controls against your requirements during the integration assessment, so the security review is explicit to your situation.
Pricing
Dust uses per-seat pricing — around €29 per user per month for Pro, with Enterprise custom. That’s transparent and accessible for small teams, but per-seat cost scales with headcount, which can add up as adoption grows. dgm’s osFoundry engagement pricing is fixed and public instead: $399 assessment and $3,999/month integration, with no per-seat fees — so cost doesn’t rise as more of your team uses it.
Self-serve workspace vs implemented orchestration
The practical difference is delivery model. Dust is a self-serve workspace your team builds agents in — great if you have people who want to build and the appetite to do it per seat. osFoundry, via dgm, is an implemented orchestration layer: dgm scopes, builds, integrates, and trains, with the explicit aim of consolidating tools. If you’d rather have it done than build it, and you value predictable non-per-seat cost and tool consolidation, that’s the distinction. The two aren’t mutually exclusive in concept, but they represent different ways to get to “agents doing real work.”
Who each is best for
Dust is the stronger choice if you want a polished, self-serve multiplayer agent workspace and have teams happy to build their own agents, accepting per-seat pricing. osFoundry is the stronger choice if you want orchestration and SaaS consolidation delivered as an implemented system with predictable, non-per-seat cost.
Which should a US company choose?
If a self-serve agent workspace with per-seat pricing fits your team, Dust is a strong, model-agnostic option. If you want orchestration plus consolidation, implemented for you with fixed cost, then osFoundry is the more direct fit. dgm assesses your goals, recommends the right path for a US business, and implements it end to end.