Convert a Docker Compose file into a POSIX sh script that runs its services as a single Podman pod.

Built for CI and test environments where you can't use docker compose or podman kube play:

• No bridge networking / netavark. Unprivileged CI containers often have a read-only /proc/sys, so netavark fails to create bridge networks. A single pod shares one network namespace with no bridge: services talk over 127.0.0.1, and names resolve via a generated /etc/hosts the script owns.
• No systemd. Podman healthchecks are normally scheduled by systemd timers. compose2pod gates startup by polling podman healthcheck run directly, so depends_on: service_healthy works without systemd.
• No heavy runtime. The core is stdlib-only — no dependencies, no compiled wheels — so it installs and runs in minimal Python images.

Requirements

compose2pod's generated scripts own /etc/hosts: they write it to a temp file and bind-mount it read-only into every container under --no-hosts (see architecture/supported-subset.md), so pod-internal name resolution works on any Podman version. host.containers.internal / host.docker.internal are not provided — add an explicit extra_hosts entry if you need them.

Install

pip install compose2pod            # core: reads compose as JSON
pip install compose2pod[yaml]      # optional: read YAML directly (adds PyYAML)

Usage

# YAML directly (needs the [yaml] extra)
compose2pod docker-compose.yml --target app --image myimage:ci > run.sh

# Or stay dependency-free by piping JSON (e.g. via yq)
yq -o=json '.' docker-compose.yml | compose2pod --target app --image myimage:ci > run.sh

sh ./run.sh

Supported compose subset

compose2pod refuses every document docker compose config refuses — a measured property, checked continuously by a differential conformance harness that runs the real Docker CLI and the real compose2pod pipeline over the same YAML. So a file that compiles is a file Docker would run; where compose2pod still refuses a form Docker accepts, it is because Podman genuinely cannot express it (each such case is documented, not guessed).

Within that boundary it covers most of what real compose files use:

• Services — image/build, command/entrypoint, environment and env_file (string and long-form {path, required, format}), volumes (short-form and long-form --mount, including the bind/volume/tmpfs option maps), tmpfs, healthcheck, depends_on (all conditions), network aliases, hostname/container_name.
• Confinement & metadata — user, working_dir, read_only, init, privileged, cap_add/cap_drop, security_opt, devices, group_add, platform, labels, annotations, pull_policy (the quoted-boolean and YAML-1.1 spellings Docker accepts, too).
• Resources — the legacy keys (mem_limit, cpus, pids_limit, ulimits, …) and the modern deploy.resources block.
• Pod-wide — dns/dns_search/dns_opt, sysctls, extra_hosts.
• Composition — same-file extends, secrets/configs, profiles.

Compose extension fields (any x--prefixed key) and YAML anchors are accepted as-is, so a top-level x-* anchor block for shared config just works. ${VAR}-style variable interpolation is left live in the generated script, resolved by its shell against the environment present when the script runs (no .env file support). See architecture/supported-subset.md for the full accept/ignore/reject matrix and planning/decisions/ for the boundary rulings.

Status

Beta. Part of the modern-python family. MIT licensed.