Skip to main content

Why RocketRide

View as Markdown

Why RocketRide

RocketRide is an open-source runtime for AI pipelines. It sits one layer below your application and your agent framework, handling the models, vector stores, observability, and orchestration that come after the code, so you can swap providers, scale workloads, and ship production AI without rebuilding your stack every time the landscape shifts.

Pipelines are portable JSON​

A pipeline is a .pipe file: plain, version-controlled JSON. The same file runs unchanged on your laptop, your servers, or RocketRide Cloud. Commit it, diff it, share it, code-review it like any other source artifact, no proprietary project format, no lock-in.

Build visually or in code​

Author pipelines on a visual drag-and-drop canvas in the VS Code extension, or define and run them from a few lines of TypeScript or Python. The two are interchangeable: build visually, then run from code, or vice versa.

A fast, production-grade core​

Pipelines execute on a multithreaded C++ engine built for throughput and reliability, not a thin wrapper around HTTP calls. You get the same execution semantics in development and production.

Batteries included, nothing locked in​

  • 50+ ready-to-use nodes: 13+ LLM providers, 8 vector databases, OCR, NER, PII anonymization, transcription, web tools, and more.
  • Swap providers freely: change an LLM or vector store by editing config, not code.
  • Observability built in: trace call trees, token usage, and memory as pipelines run.
  • Open protocols: drive pipelines over WebSocket or expose them as tools over MCP.
  • Deploy anywhere: locally, on-premises with Docker, or managed on RocketRide Cloud.
  • MIT licensed: fully open-source and OSI-compliant.

Where to go next​