ARCANA is an AI engineering workbench for developers and technical teams.
ARCANA runs from the terminal and helps developers plan, execute, inspect, and verify AI-assisted engineering work. It supports multiple model providers while keeping model credentials under the user's control. Sessions expose tool use, execution steps, permissions, cost and routing evidence, and durable records so a developer can understand what an agent did instead of treating the result as an opaque chat response.
Who ARCANA is for
ARCANA is designed for software engineers, platform teams, security-conscious developers, AI infrastructure teams, and organizations that want agentic coding workflows with explicit control and evidence. It is useful when work spans code inspection, model routing, scoped tools, session memory, terminal automation, or governed execution and the operator needs a record that can be reviewed after the action.
Developer resources
Start with the developer portal or documentation, inspect the OpenAPI specification, discover public endpoints at /api, read llms.txt and agents.md, discover the Agent Skills index, or use the sitemap to enumerate public pages.