Dust vs Writer

Dust and Writer both build agents, but they solve different layers of the same problem. Dust's differentiator is connection depth and model freedom: 70+ connectors (Slack, Notion, Google Drive, GitHub, Salesforce, Zendesk), external tools via native and remote MCP servers, and per-agent choice from 20+ models including GPT, Claude, Gemini, Mistral and DeepSeek. Its multiplayer framing shows in Pods (shared workspaces), Frames (interactive dashboards and apps), Spaces for segmenting company data by team, and a dual-layer permission model separating what an agent can access from what a user can. Triggers can be scheduled or event-driven, and self-improving skills (shipped July 2026) encode team expertise and adapt from usage. Writer's differentiator is brand and governance: it encodes your voice, terminology and standards into agents, backed by a Knowledge Graph that grounds outputs in your data. AI Studio builds and governs custom agents (governance controls broadened July 2026), Playbook Hub ships 100+ prebuilt agents, and agents handle image analysis and generation, browser automation and takeover, code execution, presentation and file generation, and Canvas editing with human-in-the-loop approvals. Writer's agents run on its own Palmyra X6 model, which the company says makes agentic work faster and more cost-effective than the prior generation — so model flexibility is Dust's story, not Writer's. If you need one agent wired into six systems with the model of your choice, that's Dust. If you need consistent, governed outputs at brand scale, that's Writer.

Both list as freemium, but they land in very different budget conversations. Dust's cost behavior is variable by design: credits are consumed based on which model an agent calls and how many connected tools it invokes, and unused credits don't roll over. That means a heavy multi-agent, multi-connector rollout can be harder to forecast month to month, and Dust itself flags that buyers wanting predictable flat-rate pricing should look elsewhere. The upside is you can steer cost by assigning cheaper models to routine agents and frontier models to the hard ones. Writer publishes freemium as its pricing type and pitches enterprise deployment, so the honest read is that real budgets are enterprise-level and tied to seats, agents, and governance needs rather than per-task credits. Its Palmyra X6 launch is explicitly framed around lower cost per agentic task than the prior generation, which matters if you're running Writer agents at marketing-department scale. The practical split: Dust suits teams that can tolerate usage-based variability and want to optimize per-agent model spend; Writer suits organizations that budget enterprise contracts and care more about predictable governance and brand consistency than shaving cost per call. Check current tiers directly — neither vendor's public facts here include dollar figures, and both should be quoted against your actual agent volume.