Get the weekly radar
One short email with the freshest products, every week.
Grounded in available product and source data
Runko is a self-hosted monorepo platform built directly on Git rather than a database or a bespoke storage layer: every change goes through a Gerrit-style refs/for/main push, gets reviewed, and lands by rebase once required checks and server-enforced merge gates pass. There are no direct pushes to main and no merged pull requests anywhere in the system — each push is scoped by affected-based CI, checked against per-project required checks declared in PROJECT.yaml, and screened by a secret scan before it can land. Path ownership lets teams assign responsibility for parts of the repository, and a Postgres index of trunk exists only as a rebuildable cache, never a second source of truth.
Humans and coding agents share the same CLI and API surface, but agents operate under stricter, server-enforced policy: each agent identity is minted per task with its own token that expires by TTL, one task is enforced to equal one workspace, and an agent can never approve its own changes, create other agents, or hold operator-level powers. Oversized changes are refused at the point of receipt, with the refusal message suggesting a split into smaller changes. According to the project, this setup is already working end-to-end and used daily by its own development team.
For an organization already running Runko, getting started needs no server setup: install the CLI, point it at the org's control-plane URL, then run runko auth signup, runko workspace create, and runko project create before looping through runko change push and runko change land. Anyone can browse a public instance read-only without an account, though joining an org for real use is invite-only. Teams that want to self-host can bring up the control plane with docker compose up, which starts runkod and Postgres with two eval users for a demo review-and-land loop, and prebuilt CLI binaries are available for Linux, macOS, and Windows, or installable via go install. The project does not state any pricing, so the cost of hosting or using Runko at scale is not specified.
Runko is a self-hosted monorepo platform built on plain Git, combining change-based review, path ownership, and rebase-based landing behind server-enforced merge gates, distributed under the Apache-2.0 license.
Yes — the project lists a web UI with stacked review, a code browser, and search, along with code search powered by Zoekt and an MCP server for programmatic access.
According to the project, releases are immutable and their changelogs are generated from the changes that landed; the same feature list also includes a CI watchdog and an operator admin panel.
According to the project, runkod, the core daemon, handles multi-org setups and an outbound mirror alongside smart-HTTP git, a receive funnel, and REST plus Connect APIs.
