Hopsworks Terminal

AI agent · Coding agent · Engineering & development by Hopsworks

Where AI Agents Go to Work

A browser-based environment for AI coding agents to drive the Hopsworks platform. It enables developers to build production ML systems and dashboards in a fraction of the time using agents like Claude Code.

Agent specs

The full spec card (context, pricing per token, modalities and more) on AntSpace. Sign up free to see the full spec card.

Launched
2026-06-15
Languages
English

Category: AI Agents

Channels
In-app (Browser-based Terminal)
Editors and tools
Hopsworks Terminal
Models
Claude Code, OpenAI Codex
Deployment
SaaS, Self-managed (Cloud or On-premises)
Languages
English
Interface
Browser-based Terminal
  • Browser-based terminal for AI coding agents
  • Native support for Claude Code and OpenAI Codex
  • Secure credential mounting from HopsFS
  • Full FTI and job lifecycle management via CLI
  • Session heartbeats for long-running agent tasks

Overview

Hopsworks Terminal is a browser-based environment designed for developers to build and deploy AI systems using coding agents. It provides a dedicated container where credentials for agents such as Claude Code and OpenAI Codex are securely mounted from HopsFS, allowing them to drive the platform natively. The terminal enables agents to manage the entire Machine Learning lifecycle, from feature engineering and training pipelines to job scheduling and dashboard creation. By exposing the Feature Store and platform CLI as a direct tool interface for LLMs, it reduces the friction between agents and data infrastructure. It includes session heartbeats to maintain long-running tasks and integrates with OpenTelemetry for full observability of agent actions.

Agent Environment and Security

  • Dedicated Container Environment: The terminal is a browser-based environment backed by a dedicated container for running AI coding agents.
  • Secure Credential Mounting: Credentials for coding agents like Claude Code and OpenAI Codex are mounted directly from HopsFS for secure platform access.
  • Session Heartbeat Management: Per-session heartbeats keep long-running agent work alive and provide clear timeout notices to prevent silent idle disconnects.
  • Native Platform Control: Agents running inside the terminal can drive the entire Hopsworks platform securely out of the box.

Platform Instrumentalisation

  • Feature Store Instrumentalisation: The Hopsworks CLI and companion skills expose the Feature Store as a direct tool interface for LLMs.
  • Job Lifecycle Management: Agents can manage the full job lifecycle including create, run, stop, status, logs, history, and scheduling.
  • FTI Lifecycle Coverage: The terminal supports the complete FTI lifecycle including feature groups, training datasets, and lineage.
  • Cluster-Wide Orchestration: Agents can drive an entire Hopsworks cluster through the dedicated hops agent command group.

Developer Productivity Tasks

  • System Prototyping: Agents can be used to build production AI systems including batch, real-time, and agentic workflows.
  • Guided Platform Intelligence: The platform provides smart suggestions for compute, arguments, and data sources to guide agent actions.

What it does

  • Manage feature groups and feature views
  • Create and schedule training datasets
  • Execute and monitor ML jobs
  • Build and deploy Apache Superset dashboards
  • Drive cluster-wide platform operations via CLI

Use cases

  • Data engineers use agents to manage the full lifecycle of feature groups and training datasets through the Hopsworks CLI.
  • Platform engineers run long-running agentic workflows for cluster management with session heartbeats to prevent idle timeouts.
  • Business analysts use coding agents to build and deploy Apache Superset dashboards from federated data sources in seconds.

Specifications

Channels
In-app (Browser-based Terminal)
Integrations
Claude Code, OpenAI Codex, HopsFS, Apache Superset, Trino, Hugging Face, MongoDB, Databricks Unity Catalog
Editors and tools
Hopsworks Terminal
Human control
Users can open the terminal to bring their own agents and monitor agent activity through session heartbeats and OpenTelemetry tracing.
Models
Claude Code, OpenAI Codex
Deployment
SaaS, Self-managed (Cloud or On-premises)
Languages
English
Interface
Browser-based Terminal
Supported Agents
Claude Code, OpenAI Codex
Credential Storage
HopsFS
Observability
OpenTelemetry tracing
CLI Group
hops agent
Session Management
Per-session heartbeats
Deployment Options
SaaS, Self-managed
Platform Version
5.0 and higher

How it works

  1. Access the Terminal: Users access the Hopsworks Terminal directly through the browser-based platform interface.
  2. Secure Agent Authentication: The environment automatically mounts necessary credentials for agents like Claude Code from the HopsFS file system.
  3. Execute Agentic Workflows: Developers use coding agents to drive the platform via the Hopsworks CLI to build feature pipelines, train models, or create dashboards.
  4. Monitor and Trace: The platform monitors agent activity with session heartbeats and provides full observability through OpenTelemetry tracing.

FAQ

What is the Hopsworks Terminal?

It is a browser-based terminal environment where developers can run coding agents like Claude Code to drive the Hopsworks platform and build AI systems.

Which AI agents are supported?

It natively supports Claude Code and OpenAI Codex by mounting their credentials directly from HopsFS.

Can it handle long-running agent tasks?

Yes, it uses per-session heartbeats to keep long-running work alive and provides clear timeout notices if a session is about to expire.

What parts of the Feature Store can agents control?

Agents can manage the full lifecycle of feature groups, feature views, training datasets, transformations, tags, and lineage.

Added by AntSpace from public sources (hopsworks.ai).

Related

More AI agents by Hopsworks