What RushDB Cloud Adds

RushDB Cloud runs the RushDB engine and adds the features below. On a self-hosted engine they are not available: the SDKs throw CloudOnlyError before sending anything, and the MCP server doesn't list the matching tools.

Feature	In RushDB Cloud	On a self-hosted engine
Managed embeddings	Create a managed index on a property; Cloud embeds every write and backfills existing records, and embeds text query searches.	Create an external index and supply vectors from your own model. See Bring Your Own Vectors.
Natural-language search	ai.search(prompt) turns a request into a SearchQuery over your schema and runs it.	Build the SearchQuery yourself, or let an agent build it from getSchemaMarkdown.
Relationship-pattern suggestions	Cloud analyzes your schema and proposes relationships to materialize, for your review.	Create relationships with attach or relationships.createMany.
Dashboard	Browse and edit records, explore the graph, build queries, manage embedding and text indexes.	Use the SDKs, the engine's native protocol, or SQL tools over the PostgreSQL wire protocol.
Projects, workspaces and team access	Separate projects, invite teammates, full-access and read-only API keys.	One engine per deployment; tenants and tokens are yours to manage.
REST API	https://api.rushdb.com/api/v1	The engine's native HTTP protocol.
Hosted MCP with OAuth	https://mcp.rushdb.com/mcp for ChatGPT, Claude.ai and other web clients.	The local MCP server in direct mode, over stdio.
Operations	RushDB runs, upgrades and backs up the engine.	You do: see Operations in the engine manual.

Everything else — records, relationships, SearchQuery, aggregations, imports, transactions, keyword, vector and hybrid search, raw openCypher — is the engine's, and works the same in both. See the SDK feature matrix.