Installing EPT¶
Three ways to run EPT — same API either way.
- EDITO service — zero install, an EDITO account is all you need.
- Docker — prebuilt images, no clone.
- Clone —
uv syncand run.
EDITO service¶
Launch the service playground from your personal space. Sign in in the UI, or point scripts at your service URL. Nothing below applies.
Docker¶
Docker is the only prerequisite. Prebuilt API and frontend images are on GHCR — no clone needed:
docker network create ept-net
docker run -d --name ept-api --network ept-net -p 8000:8000 \
-v ept_workspace:/app/workspace -v ept_data:/app/data \
ghcr.io/edito-infra/edito-publishing-toolkit-backend:latest
API on :8000, docs at /docs. No environment variables needed —
defaults are baked in.
Optional UI — run the prebuilt frontend against it (it proxies /api to
the backend):
docker run -d --name ept-frontend --network ept-net -p 3000:8080 \
-e EPT_API_UPSTREAM=http://ept-api:8000 \
ghcr.io/edito-infra/edito-publishing-toolkit-frontend:latest
UI at http://localhost:3000; sign in with your EDITO account. Pin a
version tag instead of latest to match the backend. More (compose,
logs, state): Docker usage.
Local source code¶
Prerequisite: uv — installs the Python
required by pyproject.toml itself.
git clone https://github.com/EDITO-Infra/edito-publishing-toolkit.git
cd edito-publishing-toolkit
uv sync
uv run pytest # verify
uv run uvicorn ept.api.main:app --host 0.0.0.0 --port 8000 --reload
API on :8000, docs at /docs. Next: usage. The Makefile
wraps these (make install, make test, make api-start);
api-stop/api-restart are Unix-only.
Optionally add the UI with the prebuilt frontend image:
docker run -d --name ept-frontend -p 3000:8080 \
-e EPT_API_UPSTREAM=http://host.docker.internal:8000 \
--add-host=host.docker.internal:host-gateway \
ghcr.io/edito-infra/edito-publishing-toolkit-frontend:latest
UI at http://localhost:3000 — no local Node install needed.