Enterprise integrations, operational data by design
Dataspine is for SaaS and AI product teams that need repeatable, high-stakes enterprise integrations without rebuilding pipelines for every customer. The platform provides data consistency by design and a governed access layer so your data products, connectors, and applications read and write the same canonical operational truth.
One layer: ingest, process, and deliver with exactly-once semantics, global ordering, and full history for replay, debugging, and new consumers. European deployments align with GDPR and data-sovereignty expectations where that matters to your customers.
Why teams choose Dataspine
The same story as our marketing site, in documentation terms: when standard API tools and point-to-point wiring are not enough, you get deep access to mission-critical and legacy systems, bi-directional control where your product must write back safely, and reliability and real-time delivery as part of the product contract—not an afterthought.
Pre-built, reusable connectors and a unified delivery model compound over time. You ship against one data layer; the complexity of each customer's landscape stays in the operating model and platform so integration stops being a one-off project every time.
Access patterns: your data, your interface
The same logical data product can be consumed the way your stack expects (generated clients and MCP are the usual integration paths; low-level transports are for advanced and tooling work):
| You need | How it fits the docs |
|---|---|
| MCP — AI-native, governed business context for agents and copilots | MCP on every data product |
| REST & APIs — services and app backends | Application integration |
| Event streams — real-time, ordered consumption | Client libraries and architecture |
| Webhooks | Part of the ingest/delivery model; start from data product lifecycle and Spine language |
| SQL (where your deployment offers it) | Governed, structured access for analysis; use platform context for when SQL fits your program |
What you get on the platform
- Correctness — Exactly-once processing, definitive ordering, replayable history, and traceability for trust and operations.
- Bi-directional operations — Read and write where contracts allow, with access control at the boundary.
- One flow: ingest, process, deliver — Ingest from streams, files, REST, webhooks, and more; transactional, reproducible processing; deliver through the patterns above, with generated clients and MCP so you are not hand-gluing schema drift.
- Enterprise operations — Isolated management vs data planes, observability, audit-friendly patterns, and token-based identity integration.
How this documentation is organized
The shape is close to the TypeScript documentation: a handbook for what most people do every day, reference for the platform, and a formal layer (grammar, wire) for the few who need it.
- Learn — Business and engineering context: the data products section (ingests, processor, outlets, APIs, client SDKs, control plane, auth, observability, applications, organizations, lifecycle), the Spine language handbook, and getting started / tutorials. Start here for narrative flow.
- Reference — Data product platform (same overview as in Learn, plus CLI and commands), the Spine language handbook (pipelines, runtime library chapter), runtime API libraries, and application integration (including MCP). CLI and Admin API entry points are expected to link to the matching concept pages under Data product platform where relevant.
- Resources — Spine language (specification): lexical and grammar reference through the appendix, plus internal binary serialization and transports (gRPC/HTTP). Intended for tool authors, language implementors, and deep integration, not the default reading path.
Get started
- Quick start for data engineers — CLI, your first Spine product, and validation.
- First data product tutorial — end-to-end flow with golden-style checks.
- Platform onboarding — organization-wide rollout and governance.
Outside this site: for commercial context, dataspine.ai, the platform page, and the epilot case study mirror how we talk about speed to production and operating model with customers. The company is also on LinkedIn.