OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

PenguinHarness

PenguinHarness is an open-source, local-first multi-agent platform for developing AI applications, with workflows for building, evaluating, optimizing and deploying them. It runs as a desktop app or CLI-hosted Web experience and includes tools for agent, skill and model management, evaluation and request tracing.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

PenguinHarness — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use PenguinHarness when the task is developing and improving an AI application through a multi-agent workflow, rather than simply using a ready-made agent for one end-user task.

Who it is for

  • Developers building AI or agent applications.
  • Teams or individuals who want to evaluate and iteratively optimize agent applications.
  • Users who prefer running a development workbench on a desktop, workstation or server.

Documented uses

  • Generate an AI application scaffold, code and run instructions from a natural-language description.
  • Evaluate an agent application with benchmarks and inspect requests in Trace.
  • Iterate on an application by identifying benchmark weaknesses and producing a new version.
  • Manage agents, skills and models through the Web experience.
Before you choose

Requirements & limits

The project is open-source under the Apache-2.0 license. The README documents desktop and CLI installation; desktop startup opens already signed in, while CLI setup uses a first-login link to claim the built-in admin account and set a password.

  • A model API key is required for at least one model.
  • The documented platforms are Linux, macOS and Windows 10+ on x64 or arm64.
  • The one-line installer bundles the runtime; npm installation requires Node.js 24 or later.
  • The desktop app and CLI share the ~/.penguin/data root and can be used interchangeably.
  • The Chrome extension is optional and gives agents access to tabs explicitly handed to them.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install the CLI and open the Web experience

Set up PenguinHarness on a workstation or server and access its browser-based interface.

Before you start

  • Linux, macOS or Windows 10+ on x64 or arm64.
  • A model API key for at least one model.
  • For npm installation, Node.js 24 or later; the one-line installer bundles its runtime.

Steps

  1. Install the CLI using the one-line installer, npm or an offline package.

  2. Run penguin web to serve the Web experience at http://127.0.0.1:7364.

  3. On first CLI startup, open the displayed first-login link, claim the built-in admin account and set a password.

  4. Configure a model on the in-app Models page.

Expected outcome

The local Web experience provides multi-session chat, agent, skill and model management, usage statistics, Trace and an evaluation center.

Guide limitations

  • The service requires an API key for at least one model.
  • The first-login link and password flow described here applies to CLI installs.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
11 October 2026
Original website title
Prism-Shadow/penguin-harness · Official homepage title · retained in its source language
How we review sourcesSuggest a correction