Context Bridge

Working with AI assistants across long sessions or across multiple tools means constantly re-explaining context. What did we decide last time? What's the current state of this project? Why did we go with this approach?

Context Bridge solves that by giving you a lightweight CLI that stores, retrieves, and injects structured context into new conversations — so the AI you're talking to today knows what the AI you talked to yesterday knew.

How It Works

You maintain a context file (or set of files) for a project. Before starting a new AI session, you run context-bridge inject and it prepends the relevant context to your prompt. It's simple, composable, and model-agnostic.

npm install -g context-bridge
context-bridge init        # sets up a .context/ directory
context-bridge add "we decided to use Velite over Contentlayer"
context-bridge inject      # outputs context block for your next prompt

Why This Exists

The problem is real. AI sessions are stateless. Every new conversation starts cold. This matters more as you use AI for serious work — where decisions compound, where earlier reasoning shapes later choices, and where "just re-explain it" wastes time and loses nuance.

Context Bridge is the minimum viable solution: a plain-text, version-controllable, shareable context layer that sits between you and any AI tool.

Stack

TypeScript · Node.js · Published to npm · MIT licensed