RocketRide

Overview

RocketRide is an open-source AI pipeline engine and development environment. It defines workflows as portable JSON .pipe files, built visually in VS Code or through the CLI, and executes them on a multithreaded C++ runtime. It supports LLM workflows, multi-agent orchestration, observability, SDK integration, and deployment on Docker, on-premises, or RocketRide Cloud.

Key Features

• Visual pipeline builder in VS Code with portable JSON .pipe files and real-time observability for token usage, LLM calls, latency, and execution.
• High-performance multithreaded C++ runtime designed for AI and data workload throughput.
• Pipeline nodes cover LLM providers, vector databases, OCR, NER, PII anonymization, chunking, and embedding models, and are Python-extensible.
• Multi-agent workflows with built-in CrewAI and LangChain support, shared memory across pipeline runs, and nodes callable as tools.
• TypeScript, Python, and MCP SDKs integrate pipelines into applications or expose them as callable tools for AI assistants.
• Deployment paths include Docker, local or on-premises operation, and managed RocketRide Cloud.

Use Cases

• Build RAG pipelines that retrieve relevant documents and ask an LLM questions with them in context.
• Process PDFs and images to extract text, find names and addresses, and detect personal data.
• Orchestrate multi-agent workflows where agents call tools and nodes repeatedly until a task is complete.
• Call RocketRide pipelines from Python or TypeScript applications through the SDK.

Getting Started

• Install the RocketRide extension from your IDE marketplace and open it.
• Deploy a server by choosing Local (recommended) to pull the server into the IDE, or On-Premises to run on your own hardware.
• Open examples/document-processor.pipe, start it from the source node, and send input through the pipeline.
• Install the rocketride SDK and call the running pipeline from Python or TypeScript; the extension writes connection details into the opened folder's .env.

Deployment & Requirements

• Local deployment pulls the server into the IDE without additional setup.
• On-premises deployment can run the Docker image or build from source.
• Docker deployment requires Docker to be installed.
• RocketRide Cloud is managed hosting with no infrastructure to provision.

Before You Adopt

• License: MIT. Review its terms before using, modifying, or distributing the project.
• On-premises deployment supports data residency and air-gapped operation for teams that need data and model calls to stay on their own infrastructure.
• Running the SDK outside the folder opened in the IDE makes the client fall back to RocketRide Cloud.