kigya vs Dust: knowledge workspace for big teams, autonomous co-founder for the team-of-one.
Dust is the right answer when you're an AI Operator inside a company with thousands of employees, a CISO, and a SCIM directory. We're the right answer when you're the founder of a company that doesn't have those things — yet.
Dust: Multiplayer AI workspace for human-agent collaboration. Enterprise-grade permissions and a semantic context layer.
Why teams pick Dust
- Enterprise-grade: SOC 2 Type II, GDPR, EU residency, HIPAA-ready
- Dual-layer permissions with SCIM-synced groups
- Semantic context layer over company knowledge
- 100+ production connectors + REST API + MCP
- Recent $40M raise (Axios, May 18, 2026)
Why teams pick kigya
- Autonomous-cofounder cycle, not knowledge augmentation
- Engineering agent that ships verifiable git diffs
- Per-company billing + agent budgets — built for team-of-one
- 3,000+ tools via managed OAuth (Pipedream Connect)
- Public `/live` scoreboard + revocable share receipts
Where Dust earns the credit
Dust's permission model — agent access decoupled from human access, SCIM-synced groups, admin-gated overrides — is the right shape for AI in a real enterprise. We'll converge here as we move upmarket.
The semantic context layer is a meaningful step beyond plain-RAG retrieval. Dust agents understand company knowledge, not just retrieve it.
Dust's framing of the 'AI Operator' role — the person inside a company who builds and runs agents — is the cleanest description of the new role we've seen.
Side-by-side
| Axis | Dust | kigya |
|---|---|---|
| Target buyer | Enterprises with 100+ employees. CISO-driven procurement. Per-seat pricing. | kigya winsFounders running a single company. Per-company pricing. No CISO, no SCIM, no procurement cycle. |
| Shape of the product | Multiplayer workspace where humans and agents share knowledge and tools. | kigya winsAutonomous co-founder running the company while you sleep. CEO + Eng + GTM + Finance agents on a nightly cycle. |
| Engineering work | Agents call tools and pull from connected sources. Building software is not the focus. | kigya winsPlan → Patch → Test → Ship loop in a real Linux sandbox. Diffs land in receipts. |
| Receipts | Audit logs (365-day retention) for compliance review. | kigya winsEvery shipped task carries a public, revocable receipt URL. Diff, exit codes, spend, verifier — all reproducible from kigya's audit log. |
| Tool catalogue | 100+ production connectors + REST API + MCP for proprietary systems. | 3,000+ tools via Pipedream Connect (managed OAuth). MCP bridge so any kigya tool is callable from any MCP client. |
| Permissions | Dual-layer permissions, SCIM-synced groups, admin-gated overrides. | RLS at the database level, per-company audit logs, share tokens with revocation. SCIM on the roadmap. |
| Compliance | SOC 2 Type II, GDPR, EU residency, HIPAA-ready, 365-day audit log retention. | SOC 2 on the roadmap; portability today via ZIP export of everything you own. |
| Pricing | €29/user/month (Pro). Enterprise custom. | Per-company subscription with hard agent budgets. One company = one workspace = one bill. |
When Dust is the right call
Pick Dust if you're an AI Operator inside a 200-person company with a SCIM directory, a CISO, and a need to ship agents that read your Slack and Notion safely. It's the cleanest enterprise-grade workspace for human-agent collaboration we've seen.
When kigya is the right call
Pick kigya if you're the founder of a one-to-five-person company that needs a co-founder, not a workspace. We do less in the knowledge-layer dimension and more in the autonomous-execution dimension.
Try kigya alongside Dust.
One company. One Slack channel. One per-task receipt for everything we ship.