# Harness2 — AI Coding Tool Review & Pricing

> Harness2 (Harness²) is Google Research's open-source framework for recursive agent harnessing. Model weights stay fixed while it refines the harness around coding agents such as OpenCode and Codex for each task, and it accumulates lessons in a persistent improvement harness for tasks it has not seen before.

Source: https://ai.dosa.dev/tools/harness2 · Updated Oct 19, 2026 · Category: Developer Productivity & Workflow · Published by dosa.dev

## At a glance

- **Tool**: Harness2
- **Company**: Google Research
- **Category**: [Developer Productivity & Workflow](https://ai.dosa.dev/category/productivity)
- **Pricing model**: Open Source
- **Official website**: https://github.com/google-research/harness2

## Key features

- Task-level Propose-Probe-Compose refinement without ground-truth feedback
- Persistent improvement harness that keeps adopted harnesses and lessons
- Eight editable components: system prompt, guardrails, memory, tool guidance, subagents, skills, scripts and plugins
- Parallel and sequential recursion modes with a configurable budget
- OpenCode and Codex executors
- JobBench, LAB and WorkBuddy-Bench setup, judging and reporting CLIs

## Pricing

**Model**: Open Source

Free and open source under Apache-2.0. The local demo needs no model credentials; benchmark runs call Vertex AI models and incur your own API costs.

## Who it is for

- **Best for**: Researchers and agent builders who want to automatically tune prompts, skills, tools and guardrails around coding agents and measure the gains on benchmarks.
- **Not ideal for**: Developers looking for a ready-to-use coding assistant; it is a Linux and Python 3.12 research framework that needs Vertex AI setup for model-backed runs.

## Verdict

A notable research release that treats the agent harness, not the model, as the thing to optimize, with practical OpenCode and Codex integration for experimentation.

## Recent updates

Open-sourced by Google Research on October 8, 2026 with an accompanying arXiv paper; the repository was updated on October 14, 2026 and reached the Hacker News front page on October 19, 2026.

## Frequently asked questions

### Is Harness2 free?

Yes - Harness2 is Open Source. Free and open source under Apache-2.0. The local demo needs no model credentials; benchmark runs call Vertex AI models and incur your own API costs.

### Is Harness2 open source?

Yes - Harness2 is open source. Free and open source under Apache-2.0. The local demo needs no model credentials; benchmark runs call Vertex AI models and incur your own API costs.

### Who is Harness2 best for?

Researchers and agent builders who want to automatically tune prompts, skills, tools and guardrails around coding agents and measure the gains on benchmarks.

### Who is Harness2 not ideal for?

Developers looking for a ready-to-use coding assistant; it is a Linux and Python 3.12 research framework that needs Vertex AI setup for model-backed runs.

### What are the best Harness2 alternatives?

The closest Harness2 alternatives on ai.dosa.dev are Pieces for Developers, agenttrace, Vibe Kanban - all listed under Developer Productivity & Workflow.

### Who makes Harness2?

Harness2 is developed by Google Research. It is listed in the Developer Productivity & Workflow category on ai.dosa.dev.

## Top alternatives

- [Pieces for Developers](https://ai.dosa.dev/tools/pieces-for-developers) (Freemium) — Developers and teams who need to manage disparate workflow context, revisit past coding decisions, and maintain consistent documentation across multiple tools and projects.
- [agenttrace](https://ai.dosa.dev/tools/agenttrace) (Open Source) — Developers who need a privacy-conscious way to audit, monitor, and diagnose the behavior, costs, and performance of various AI coding agents locally.
- [Vibe Kanban](https://ai.dosa.dev/tools/vibe-kanban) (Free) — Solo developers and individual power users who need to run multiple AI coding agents in parallel while managing separate git worktrees and tasks locally.
- [ai-memory](https://ai.dosa.dev/tools/ai-memory) (Open Source) — Developers who frequently switch between multiple AI coding agents or machines and want a unified, persistent knowledge base that survives session ends and tool switches.
- [smolvm](https://ai.dosa.dev/tools/smolvm) (Open Source) — Running untrusted code, powering coding agents with persistent/resettable workspaces, and deploying portable VM-based workloads that require a real Linux kernel.
- [coop](https://ai.dosa.dev/tools/coop) (Open Source) — Security-conscious developers who want to run autonomous coding agents on real projects without risking host machine compromise.
- Full alternatives guide: https://ai.dosa.dev/tools/harness2/alternatives