Hive
Ivan Kuznetsov · Coding
Hive is an open-source agent workflow engine that turns a rough one-line idea into a merge-ready pull request. It runs work as a folder-as-agent pipeline, orchestrating Claude Code, Codex CLI, and Pi in the background. So instead of babysitting terminals, you capture an idea, answer a few questions in a doc, and let the daemon advance tasks while you get on with your day.
About Hive
What Is Hive
Hive is an open-source coding agent pipeline for developers who already use AI coding agents and want to run them as a pipeline instead of one prompt at a time. Each task lives in a folder, and the daemon moves that folder through nine stages that take it from a captured idea to a finished, archived task: capture, brainstorm, plan, execute, open-pr, review, artifacts, finalize, and done. It's free. The license is MIT.
The main problem it solves is attention. A single AI coding agent still needs you to prompt it, watch it fail, and restart it. Hive hands that loop to a background process. The live TUI only flags the task that needs an answer. Your actual job shrinks to answering a question in your editor and closing the laptop. That's the whole appeal for anyone chasing developer productivity.
The biggest limitation is the audience. Hive assumes you're comfortable in a terminal, editing files in vim, and configuring agent CLIs yourself. There's no point-and-click interface. If you want a hosted, browser-based coding assistant, this isn't it. Not close.
Getting Started
- Install the Hive CLI and run
hive initinside your project to create the stage folders and config. - Point each stage at an agent CLI you trust, such as Claude Code by default, or Codex and Pi for specific steps.
- Capture a rough idea in the inbox, one sentence is enough, then let the daemon pick it up.
- Answer the brainstorm questions in your editor when the TUI flags the task.
- Archive the finished task once Hive has pushed the branch and marked the pull request ready.
Product Information
A quick look at Hive's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Solo developers who already drive Claude Code or Codex CLI and want several tasks running at once without watching each one.
- Small engineering teams that want a shared, readable trail of what an agent did at every stage.
- Power users who live in a terminal and don't mind editing markdown files by hand.
Tasks
- Turning a one-line feature idea into a reviewed draft pull request without manual prompting at each step.
- Running autonomous repo patrol to map feature slices, review them, and open PRs for the ones that pass.
- Keeping existing pull requests green and mergeable while main keeps moving.
- Capturing and refining work from a phone through the Telegram bot.
Scenarios
- Kicking off several coding tasks before a meeting and letting them advance in parallel while you're away.
- A maintainer who wants automated review triage and CI fixes on incoming changes.
- Experimenting with your own multi-step workflow, such as research or writing, on the same engine.
Key features
Folder-as-Agent Pipeline
Every task is a folder, and its location on disk is its state. Moving it to the next stage is the approval gesture, and each stage leaves a markdown artefact the next stage can trust. So what does that buy you? No database and no hidden state. You can read, edit, or hand any artefact to another agent.
Multi-Agent Orchestration
Stages run on configurable agent CLIs, with Claude Code as the default and Codex or Pi available for specific jobs. Each step can use the model you trust for that kind of work, rather than forcing one agent through the whole pipeline. That flexibility is the point.
Async Parallel Queue and Daemon
A background daemon advances many tasks at once. The live TUI flags the single task that needs your input. You answer in your editor, then go back to what you were doing. That's what makes Hive a pipeline rather than a chat window.
Patrol
Opt-in autonomous repo patrol maps feature slices, reviews them, validates fixes, and opens pull requests for the ones that pass. It's a way to get work moving on a repo without starting each individual task by hand, which matters once a codebase grows past the point where you can track everything mentally.
Babysitter
Babysitter keeps open pull requests green and mergeable. It runs bounded repair attempts and auto-rebases behind a moving main branch, then hands off to you when it gets stuck, so a branch that drifts behind main doesn't quietly rot while you're focused on something else entirely.
Telegram Bot
You can capture ideas, check the queue, and approve work from your phone. Photos, documents, and transcribed voice notes are included, which is useful when an idea hits you away from your desk.
Custom Workflows
The engine is generic. Idea to PR is the flagship coding workflow, but Hive also ships a content pipeline. You can author your own per-project pipeline in YAML for writing, research, or triage.
Pros and cons
Pros
- Free and open source under the MIT license, so there's no subscription to justify.
- Full transparency, since every artefact is a markdown file you can read and edit.
- Runs several coding tasks in parallel instead of one prompt at a time.
- Works with multiple agent CLIs, so you aren't locked into a single model or vendor.
- Phone access through the Telegram bot covers idea capture and approvals on the move.
Cons
- Terminal-only with a power-user TUI, so there's a real learning curve if you're not comfortable in a shell.
- You have to configure and pay for your own agent CLIs separately, which adds setup work.
- Built-in workflows cover coding and content, so anything else means authoring YAML yourself.
Frequently asked questions
It's an agent workflow engine that turns a rough idea into a merge-ready pull request. Rather than prompting an AI coding agent step by step, you drop a one-line idea into a folder and let a background daemon carry it through planning, execution, review, and shipping.
Related content
Explore related tools, skills, and articles for Hive.
Hive Alternatives
Forefront
Forefront · CodingForefront is a web platform for building with open-source AI. It lets you fine-tune leading open-source language models on your own data, evaluate how they perform, and run them through an API or export them to host yourself. Developers who want the convenience of a closed-source platform but insist on owning their models and data are the target audience here.
Startkit
StartKit.AI · CodingStartkit is a boilerplate for building AI SaaS and AI wrapper products. Think of it as an AI startup boilerplate with the boring parts already wired up: authentication, Stripe and Lemon Squeezy payments, usage limits, transactional email, and an AI API starter that talks to OpenAI, Anthropic, Groq, or Llama. You clone the repo, set your price, and start on the part of your product that people actually pay for. It's Next.js under React and Tailwind, so most of the boilerplate code already feels familiar.
Testim
Tricentis · CodingTestim is an AI-powered test automation platform for building and running end-to-end tests across web, mobile, and Salesforce applications. It leans on machine learning to keep tests stable when an interface changes, so teams spend less time fixing broken selectors. Not bad for an automated testing tool you can start using today. You create tests by recording actions in a browser, then optionally add JavaScript when you need more control. It's a solid pick for busy QA teams.