rocketride-org/rocketride-server

RocketRide Server – Open‑source AI pipeline engine

What it is – RocketRide is an “AI Development Environment” (AIDE) that turns your regular IDE (VS Code, etc.) into a visual builder and runtime for AI/ML workflows. Pipelines are defined in a portable JSON (*.pipe) format, edited with a drag‑and‑drop canvas inside the IDE, and executed by a high‑throughput multithreaded C++ engine. The same pipeline can run on RocketRide Cloud (hosted SaaS) or on‑premises (Docker, bare‑metal, or a local process) with zero code changes.

Key features

Feature	What you get
Visual pipeline builder	Build, connect and configure >100 nodes (LLM providers, vector DBs, OCR, NER, chunking, etc.) directly in VS Code.
C++ runtime	Native multithreaded engine designed for low‑latency, high‑throughput AI workloads.
Multi‑agent support	Built‑in CrewAI and LangChain integration; agents can be chained and share memory.
SDKs	Python and TypeScript client libraries let you invoke pipelines from any app with a few lines of code.
Observability	Real‑time tracing of token usage, latency, memory, call trees, etc., from the IDE.
Zero‑ops / self‑hosted	One‑click Docker image (ghcr.io/rocketride-org/rocketride-engine) or local binary; also a managed cloud service.
Extensible nodes	Nodes are Python‑extensible; you can publish custom nodes to the marketplace.

Technology stack

• Core runtime: C++ (multithreaded, high‑performance)
• IDE integration: VS Code extension (Webview UI)
• SDKs: Python (rocketride on PyPI) and TypeScript (rocketride on npm)
• Containerisation: Docker image, Helm chart for cluster deployment
• Transport: WebSocket endpoint (ws://localhost:5565 locally) or HTTPS for the cloud service

Typical workflow

1. Install the RocketRide VS Code extension.
2. Create a *.pipe file – the UI shows a canvas where you drag nodes (e.g., a webhook source → LLM → vector DB → output).
3. Configure each node (model provider API keys, DB connection strings, etc.).
4. Press ▶ in the canvas to run the pipeline locally, or point the client to ROCKETRIDE_URI=ws://localhost:5565 (self‑hosted) or https://api.rocketride.ai (cloud) with an auth token.
5. Use the Python/TS SDK to call the pipeline from your application, or expose it as a tool for a coding‑assistant agent.

Getting started / installation

• VS Code extension – search “RocketRide” in the marketplace or download from Open VSX.
• Run the engine locally – either let the extension download the binary for you, or manually pull the Docker image:

  docker pull ghcr.io/rocketride-org/rocketride-engine:latest
  docker run -d --name rocketride-engine -p 5565:5565 ghcr.io/rocketride-org/rocketride-engine:latest
  

• Cloud – sign up at https://cloud.rocketride.ai, obtain an API token, and set:

  export ROCKETRIDE_URI=https://api.rocketride.ai
  export ROCKETRIDE_AUTH=your-api-token
  

• SDK usage (Python example):

  from rocketride import RocketRideClient
  client = RocketRideClient(uri="ws://localhost:5565")
  result = client.run_pipe("my_pipeline.pipe", input={"text": "Hello"})
  print(result)
  

License & community

• The server runtime and SDKs are released under the MIT license – fully open source and OSI‑approved.
• Development is coordinated on GitHub (issues, PRs, CONTRIBUTING guide) and a Discord community for support and collaboration.
• CI badge shows automated testing; releases are versioned (e.g., server-vX.Y.Z).

Who should use it

• Data‑science or ML teams that need a reproducible, version‑controlled way to stitch together LLM calls, vector stores, and other AI components.
• Companies that want to keep data residency on‑premises while still benefiting from a visual pipeline builder.
• Developers building AI‑augmented applications who prefer a single, language‑agnostic runtime instead of wiring many separate services.

---

RocketRide aims to be the “engine” beneath AI applications, handling everything from orchestration to observability, while letting you stay inside the editor you already love.