Quick Start

This guide creates and runs a minimal CoreMind project. CoreMind currently requires Node.js 22.19 or later. Python embedding requires Python 3.10 or later.

1. Choose your language

When creating a project, choose TypeScript, JavaScript, or Python. TypeScript is recommended for first-time users because configuration and SDK calls receive the strongest editor feedback.

2. Install and create a project

coremind-cli@1.0.1 is the target package for this release line; confirm public availability on the live npm page before installation.

npm install coremind-ai@1.0.1 coremind-cli@1.0.1
npx coremind-cli@1.0.1 providers
npx coremind-cli@1.0.1 create my-agent --provider alibaba-model-studio
cd my-agent

Select a template, language, provider, and permission mode when prompted. Non-interactive scripts must provide --provider. The generated project contains configuration, environment examples, tests, and local operating instructions.

3. Configure credentials

Copy .env.example to .env and add only the credential required by your selected provider. Never commit .env.

DASHSCOPE_API_KEY=replace_me

Credentials belong in environment variables. An embedding host may instead supply a SecretRef resolver; sensitive Headers, including common API-key and token aliases, cannot use plaintext literals.

Provider support does not imply live certification. Check the provider matrix before production evaluation.

4. Validate before running

npx coremind doctor
npx coremind check
npx coremind run coremind.yaml --dry-run

doctor checks the environment, check validates the project contract, and --dry-run resolves configuration without sending a model request.

5. Run interactively

npx coremind chat coremind.yaml

Start with the ask permission mode. Review every requested operation until you understand the tool and data boundaries.

Next steps

• Learn the configuration model.
• Add reusable instructions through Skills.
• Define quality gates.
• Explore the complete CLI reference.