Which AI models and providers does hive_review support?

It supports OpenAI Codex, Anthropic Claude, Google Gemini, xAI Grok, OpenCode (including DeepSeek models), and Moonshot Kimi. Both login and API-key authentication are available for most providers; OpenCode requires API-key auth in hive_review.

Can I use hive_review on Windows?

Yes, but only through Windows Subsystem for Linux (WSL). Native Windows terminals such as PowerShell and cmd.exe are intentionally not supported.

What happens if reviewers don't all pass after the maximum rounds?

If max rounds are reached (default: 5) and the human does not continue, the session status becomes max_rounds_reached. In interactive mode, you can type 'continue review' to add one more round.

Large Language ModelsAI Coding AssistantsHacker NewsJul 24, 2026Jul 24, 2026, 19:01 JST

Hive_review: Ruby CLI for multi-agent AI code review

Send on LINE

3 Key Points

What happened
Hive_review is a Ruby command-line tool that automates code review by running an implementor agent to write or fix code, then sending it to one or more reviewer agents, looping until all reviewers pass. It supports OpenAI Codex, Claude, Gemini, Grok, OpenCode (including DeepSeek), and Kimi as providers, with both login and API-key authentication.
Why it matters
Development teams can automate the code review cycle without manual handoffs—reviewers provide feedback, the implementor fixes the code, and the loop continues up to a configurable maximum of 5 rounds (default). Session state stays local to the repository, not in global provider homes, reducing setup friction and keeping review artifacts out of user-wide Codex, Claude, Gemini, Grok, OpenCode, or Kimi config.
What to watch
The tool runs on macOS, Linux, and Windows (via WSL); interactive mode allows you to resume existing review sessions or continue beyond the maximum rounds if reviewers still request changes. DeepSeek models are available through the OpenCode wrapper by configuring DEEPSEEK_API_KEY.

Ask the AI about this article →

Summaries like this, in your inbox every morning.

Get news like this every morning — free

Context & Analysis

Hive_review addresses a common bottleneck in code-driven workflows: the back-and-forth between implementation and review. By automating the orchestration of multiple AI agents across different providers, it lets teams define a single task (via a prompt or Git diff) and have it automatically cycle through fix-and-review iterations without manual coordination. The tool's design prioritizes isolation—all runtime state (provider homes, transcripts, session IDs, logs) stays within the repository under .hive_review_state, avoiding pollution of user-global provider config directories.

The architecture is flexible: users can choose one implementor and one or more reviewers from a JSON config file, supporting newly released models without code changes. Authentication is handled both via login (interactive) and API-key (unattended), and the tool includes helpers to install missing provider CLIs and verify auth before running. For teams using DeepSeek, the OpenCode wrapper eliminates the need for a separate provider integration, reducing setup friction.

FAQ

Which AI models and providers does hive_review support?: It supports OpenAI Codex, Anthropic Claude, Google Gemini, xAI Grok, OpenCode (including DeepSeek models), and Moonshot Kimi. Both login and API-key authentication are available for most providers; OpenCode requires API-key auth in hive_review.
Can I use hive_review on Windows?: Yes, but only through Windows Subsystem for Linux (WSL). Native Windows terminals such as PowerShell and cmd.exe are intentionally not supported.
What happens if reviewers don't all pass after the maximum rounds?: If max rounds are reached (default: 5) and the human does not continue, the session status becomes max_rounds_reached. In interactive mode, you can type 'continue review' to add one more round.

Hacker NewsRead Original Article

Send on LINE

Get the latest Large Language Models news every morning

For example, today's edition would include:

Aron launches with $8 million to automate procurement RFQsSiliconANGLE AI · 38m ago
Talkdesk's Andrade: 98% deploy AI, only 15% orchestrateSiliconANGLE AI · 38m ago
Google DeepMind: 100 AI agents split 24-14 over cheatingMITテクノロジーレビュー · 38m ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Subscribe free →

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Aron launches with $8 million to automate procurement RFQs

Aron Inc. launched with $8 million raised across two rounds

NEWSiliconANGLE AI38m ago

Talkdesk's Andrade: 98% deploy AI, only 15% orchestrate

Pedro Andrade, Talkdesk's VP of AI and generative AI business specialist, told theCUBE that CXA is "an operati…

NEWSiliconANGLE AI38m ago

Google DeepMind: 100 AI agents split 24-14 over cheating

In a Google DeepMind experiment, 100 Gemini 3.1 Pro agents were told to act as top math researchers and solve…

NEWMITテクノロジーレビュー38m ago

71.5% use generative AI in a separate screen, TechTouch survey finds

TechTouch surveyed 319 people overseeing generative AI at firms with over 1,000 employees

NEWITmedia AI＋38m ago

Microsoft, OpenAI back AI slowdown, chip stocks fall

Microsoft joined OpenAI, Anthropic, and SpaceXAI on Monday in endorsing a more prudent approach to frontier AI…

NEWSemafor Tech38m ago

iLands AI agents flood Mastodon, Bluesky, X with spam

AI agents from startup iLands, including one called Ren, sent unsolicited messages to Mastodon admins and writ…

NEWArs Technica AI38m ago

Next articleUS targets Chinese humanoid robots in new tech rivalry move