Skip to content

IVCAP CLI

ivcap is the official command-line interface for interacting with an IVCAP deployment. It can be used directly by humans at a terminal, and is also used by AI agents via its built-in MCP server.


What is IVCAP?

IVCAP (Integrated Virtual Collaborative Analysis Platform) is a managed, provenance-aware platform for running, building, and orchestrating analytic services and AI agents. It is designed to help researchers investigate their domains and derive new insights by:

  • collecting, processing and analysing multi-modal and multi-scale data,
  • tracking provenance automatically — recording which service processed which inputs to produce which outputs, and
  • fostering collaboration across disciplines by making data lineage transparent and reproducible.

IVCAP exposes a REST API. This CLI wraps that API into convenient shell commands and also acts as a local MCP server so that AI agents can interact with the platform programmatically.


Core concepts

Each row links to the CLI command that manages that resource.

Concept Description Command
Service A packaged, executable capability. Services are registered by developers and invoked by any authorised user. service
Job A single execution of a service. Jobs progress through states (pending → executing → succeeded/failed) and emit events. job
Artifact Immutable blob-like data (images, JSON, tar'd datasets) produced and consumed by jobs. artifact
Collection A named group of artifacts or resources, useful for organising experiment outputs. collection
Data Fabric A metadata graph storing aspects — structured records attached to any entity (job, artifact, service, …). Used for provenance and discovery. datafabric
Queue An ordered message channel for asynchronous communication between services. queue
Secret A named credential stored securely in the platform for use by services at runtime. secret
Package A container image in the platform's registry — the execution unit behind a service. package
Context A named deployment endpoint and credentials for the CLI to connect to an IVCAP installation. context
Nextflow pipeline A Nextflow DSL2 pipeline deployed as an IVCAP service for scalable scientific workflows. nextflow

Agent support

Concept Description Command
MCP server Exposes all platform operations as Model Context Protocol tools for AI agents. mcp
Skill playbook Embedded best-practice workflow guides that agents can read at runtime. skills
Agent context Embedded operational guidance (safe defaults, URN conventions, etc.) for AI agents. agent-context

New to IVCAP? See the Getting Started guide for installation, context setup, and your first job.