Developer overview
DashQ delivers governed property intelligence through external AI services, embedded experiences, scoped JSON APIs, and direct warehouse access. Choose each capability independently; inference, MCP, embeds, APIs, and BigQuery have distinct credentials and authorization boundaries.
Choose an integration
| Need | Start here | Delivery model |
|---|---|---|
| A complete dashboard or conversational experience | Embedded experiences | Your backend requests a short-lived URL; the browser renders it. |
| DashQ data in your own components | Data APIs | Your backend creates a scoped session and returns only the data your UI needs. |
| Governed warehouse analysis from an AI client | DashQ Data MCP | The user signs in with OAuth; tools are read-only and account-scoped. |
| Claude Cowork or approved external model execution | Dashnet inference | The managed Cowork pack combines the production inference gateway, OAuth MCP, and controlled output scope. |
| Definitions, formulas, reporting fields, SQL logic, and lineage | Data Dictionary | Search the complete generated public metadata snapshot. |
| Direct SQL for an approved enterprise workflow | Direct warehouse connection | DashQ provisions project, dataset, identity, approved public egress networks, and billing scope. |
Available capabilities
- Six foundation dashboard surfaces: Portfolio Performance, Lead-to-Lease, Tour Performance, Application Operations, Marketing / Acquisition, and Communication Performance.
- Dashboard and Assistant embeds with account, property, date, dashboard, and user scope.
- Dashboard and widget discovery from the same scoped session model.
- JSON data for native charts, tables, maps, cards, and paginated detail views.
- A public dictionary with 184 curated and model-derived terms, plus reporting tables/columns, formulas, SQL expressions, grain, keys, nullability, cadence, and lineage.
- Dashboard reports and generated insights inside the protected embed experience.
- OAuth-protected, read-only warehouse analysis through DashQ Data MCP.
- Anthropic-compatible model execution through
https://inference.dashnet.ai, with client-specific model and quota policy. - Provisioned BigQuery access for SQL clients, notebooks, BI systems, and locally operated MCP tools.
Trust boundary
- Authenticate the user in your application.
- Resolve the user's account, property, role, and data entitlements on your backend.
- Call the DashQ URL assigned to your environment with a backend-only credential.
- Request only the dashboards, widgets, dates, and properties the user may access.
- Return a short-lived embed URL or a minimal, allowlisted data response to the browser.
Never place embed, API, or inference keys in browser code, URLs, logs, prompts, or published configuration. Never reuse an inference key as an MCP or warehouse credential, or mint scope from untrusted request fields. Read the security requirements before implementation and use the API reference for the analytics API contract.