QUANTUMCODER FOR YOUR TERMINAL

Your models.
Your project.
One CLI.

Run a coding agent from the folder you are working in. QuantumCoder ModelRouter handles model requests; the CLI reads, edits, and checks files locally.

1. Install from the repository

From the QuantumCoder repository root, install the local CLI package globally.

npm install --global ./packages/cli

2. Connect your account

Paste an API key created in your QuantumCoder dashboard. The key prompt is masked.

quantumcoder auth

3. Start inside a project

cd path\to\your-project
quantumcoder

Build mode is automatic: describe the task and let the agent edit files and run checks. Press Tab for read-only plan mode. Use /mode confirm when you want approval before each write or command.

Designed for the workflow

Project context

Explore files from the current working directory. Local reads and writes are scoped to that project.

Plan or build

Ask for an implementation plan, then switch to build to have the agent make changes and run checks.

Your models

Choose among routes actually enabled in the QuantumCoder catalog. Models need OpenAI-compatible tool calling for coding actions.

Need to connect a self-hosted instance?

Set QUANTUMCODER_API_URL to the HTTPS API origin before running quantumcoder auth.