Dust

Dust gives companies a shared workspace where staff build AI agents and then use them alongside colleagues. Each agent can combine company knowledge, skills and tools. Agents connect to company data through connectors for Slack, Notion, GitHub, Google Drive and others, and can reach any further tool through MCP servers. Agents can also run multi-step workflows on schedules or triggers. Dust does not lock you into one model: it offers more than 20, including GPT, Claude, Gemini, Mistral and DeepSeek, and no model is reserved for a higher plan. Its permissions follow those of the connected source systems. Pricing is per seat and based on credits, quoted in euros excluding VAT. A Free seat comes with 500 lifetime credits. Pro costs 30 euros a month, or 24 a month on annual billing, with 8,000 credits per month. Max costs 150 euros a month, or 120 on annual billing, with 40,000 credits. Unused credits do not roll over, and admins can enable capped overage. The Business plan is meant for teams of up to 100 people and limits each workspace to 3 connectors, 5 remote MCP servers and 5 Spaces. Enterprise has custom pricing and adds pooled credits, unlimited connectors, SCIM, audit logs, single-tenant deployment and usage analytics. Dust suits tech-forward mid-market companies and enterprises, especially European ones that care about EU data residency. Named customers include Qonto, Alan, PayFit, Datadog and 1Password. The main limitation is that answer quality depends on how clean and up to date the connected sources are. The Business plan's three-connector cap also pushes larger rollouts towards Enterprise pricing, which requires a sales conversation.