Docs / Quickstart

Start with one local setup path.

HiveRunner is designed to be cloned, configured, and run on your machine first. Optional runtimes can be connected after the core app is alive.

Quickstart

Copy one setup block. Run the operator loop locally.

HiveRunner is not presented as a hosted SaaS. The first path is a local command center with local auth, local workspace files, and a local execution loop you can inspect.

01Open http://localhost:3010 after the dev server starts.

02Use the local continue path. No hosted account is required for the default local mode.

03Port 3010 is observer-only by default (MC_ENGINE_TICK=off). Autonomous execution belongs to a stable lane such as 3001 or another explicit isolated execution lane.

04Keep local-single-user mode on the local machine. Do not expose it to a LAN or the public internet.

local setup

$ git clone https://github.com/timharris707/hiverunner.git hive-runner$ cd hive-runner$ cp .env.example .env.local$ npm install$ npm run dev

One paste starts the local-first path. Website development uses pnpm separately; the public HiveRunner app quickstart uses npm.

How it works

A practical loop for keeping agent teams under control.

The sequence is intentionally simple: define the goal, plan the sprint, assign tasks, track runs, review before close.

Define the goal

Turn an objective into a visible goal with acceptance checks before agents start moving.

Plan the sprint

HiveRunner turns the goal into a scoped sprint — a complete unit of work, not a flat to-do list.

Assign tasks

Split the sprint into owned tasks and route each one to the right runner.

Track runs

Watch heartbeats, comments, artifacts, and status changes as execution unfolds.

Review before close

Keep human review in front of risky commands, plan changes, and final closure.

Requirements

Required and optional are kept separate.

Node.js 20 or 22 for the public app path.

npm for the HiveRunner product repo quickstart.

A local workspace directory for files, artifacts, and run context.

Optional provider keys or CLI runtimes only when you choose those integrations.

Local-first notes

Straight answers before you install.

Is a hosted account required?

No. The default public path is local-first and uses the local owner flow.

Do I need provider API keys immediately?

No. Provider keys and local CLIs are optional until you choose to run a specific integration.

What should I expect after boot?

You should reach the local app, create or open a workspace, and inspect goals, tasks, agents, hives, memory, and runtime readiness from the operator surface.

Is this a hosted multi-user service?

Not yet. HiveRunner is currently presented as a serious local-first operator tool. Hosted multi-user deployment is separate future work.