Skip to main content

Functional Groups Overview

The Nestbox AI CLI provides a set of commands for managing projects, compute infrastructure, agents, documents, images, document processing pipelines, and AI-assisted configuration generation. These commands are grouped into the following categories based on their function.

This page gives an overview of each category. You can follow the links to learn more about the available commands and their usage.

Projects​

Project commands allow you to list, configure, and switch between workspaces. You can define a default project, add project aliases, and store this configuration locally so that project-specific settings are tied to your working directory.

Compute​

Compute commands let you create and manage the compute instances that host your agents. You can list existing instances, create new ones, and delete instances that are no longer needed. Each instance is tied to a project.

Agents​

Agent commands cover the full lifecycle of AI agents — from creating and configuring deployment parameters to uploading your code and removing agents. These commands support single agent operations as well as bulk operations via the nestbox-agents.yaml manifest.

Documents​

Document commands allow you to manage collections of structured or unstructured data used by your agents. You can create and update collections, upload individual documents or files, retrieve content, and perform semantic or filtered searches. This enables integration of knowledge and reference data directly into your agent workflows.

Images​

Image commands let you list the available compute images that can be used when deploying instances.

Document Processing​

Document processing (doc-proc) commands provide a comprehensive pipeline for processing documents with AI. This includes uploading documents, managing processing profiles, monitoring jobs, running evaluations, executing batch queries, and configuring webhooks for event notifications.

Generate​

Generate commands use AI (Anthropic Claude or OpenAI GPT) to produce validated configuration files from plain-English instruction files. This includes scaffolding new projects, generating document processing pipeline configurations, and creating report composer configurations.