Developer portal
Where to start building on dcupl and where to look things up: the quickstarts, the REST API reference, the entry points for AI agents, API keys, a sandbox and support.
npm install -g @dcupl/cliThe CLI and the SDK are free on npm.
Get started
Install the SDK or the CLI, then open your project in the Console.
SDK quick start
Install @dcupl/coreand build a filterable product list: a query by category, no backend required.npm install @dcupl/core
CLI installation
Install the dcupl CLI once and run it from any directory, or on demand with npx. Installing and scaffolding need no credentials.npm install -g @dcupl/cli
dcupl Console
Where a project's files live: models, data, the loader configuration and scripts. Deploy workflows and REST API instances from it. console.dcupl.com
API reference
A REST API (RAPI) instance is deployed from your data model, with an OpenAPI document and an API key.
REST API (RAPI)
Templates, deploys and endpoints: query, facet, aggregate, pivot, suggest and metadata for every model, scripted endpoints and a status route. OpenAPI description
The OpenAPI 3 description of the dcupl REST API that every deployed instance serves, for a generated client or an agent. /openapi.json
Swagger UI of an instance
Every instance serves its own Swagger UI to try requests, and its OpenAPI document at /swagger-json. Neither needs the API key.{runner-url}/instance/{instanceKey}/swagger
For AI agents
The entry points for coding agents and MCP clients, machine-readable where they can be.
MCP endpoint
Where it is turned on, a REST API instance can be an MCP server: read-only tools for your deployed models, with the instance's API key. Part of a paid plan. POST {runner-url}/instance/{instanceKey}/mcp
MCP server card
A machine-readable description of the MCP endpoint an instance can have where it is turned on, for MCP clients that discover servers. /.well-known/mcp/server-card.json
llms.txt
The index of dcupl.com for AI agents: every page with its description, and the way to the docs. /llms.txt
Docs index for agents
Every documentation page with a description, plus per-section full-text files that fit one agent context. docs.dcupl.com/llms.txt
Agent setup
Give Claude Code, Cursor, Codex or Copilot the real data model: install the CLI and the dcupl skill, then check that the agent can drive it. Agent-ready CLI
How we made the dcupl CLI something an agent can drive: --jsonanswers, a stage and build lifecycle, schema discovery and a Claude Code skill.
API keys
Two kinds of keys: the project's keys from the Console, and the key of a REST API template.
Project keys
In the Console under Project Settings → API Keys: CLI read/write, SDK readonly and scoped keys. A browser app gets the Browser (read-only) key. REST API key
Set in the API template, one key per instance, sent as the api-keyheader or query parameter. A template without one deploys an open instance.CLI credentials
Run dcupl config setonce per workspace: it writes the project ID to the workspace config and your API key todcupl.secrets.jsonor.env.dcupl config set
Sandbox and support
Try dcupl in the browser before you install anything, and reach us when something breaks.