Marti844/SaaS-Bench-docker

SaaS-Bench Docker Images

Docker image archives for the SaaS-Bench benchmark — a suite of 23 self-hosted SaaS applications used to evaluate computer-use LLM agents on real, multi-step business workflows.

This repository hosts the prebuilt .tar images (≈ 52 GB total) so you can reproduce the benchmark environment without rebuilding each app from source. The eval harness, task definitions, and verifiers live in the main SaaS-Bench repository.

Paper: SaaS-Bench: Can Computer-Use Agents Leverage Real-World SaaS to Solve Professional Workflows?

Overview

SaaS-Bench evaluates browser-driving LLM agents on 106 task instances across 6 domains, running on 23 self-hosted SaaS applications. Each task asks the agent to complete a multi-step workflow (e.g. create a purchase order, configure a project board, schedule a patient visit); a per-task verify.py script then inspects the running application's state (DB rows, API responses, filesystem) and returns a pass/fail.

Track	Domain	Tasks	Representative apps
uni-m	BOF	15	Twenty, Bigcapital, HRMS, Pretix
uni-m	HA	16	OpenEMR, OnlyOffice, OpnForm
uni-m	SEPM	31	Baserow, OpenProject, code-server, Metabase
uni-m	TCDW	12	OnlyOffice, Mattermost, RoundcubeMail, ownCloud
multi-m	AASC	12	Grocy, farmOS, Recipya, e-label
multi-m	IMC	20	SiYuan, Watcharr, BookLore, PhotoPrism, MediaCMS

Domains: BOF = Business Operations & Finance · HA = Healthcare & Administration · SEPM = Software Eng. & Project Mgmt. · TCDW = Team Comms & Document Workflows · AASC = Agriculture, Authoring & Supply Chain · IMC = Information Mgmt. & Creative.

Contents

23 Docker image archives (mw-*.tar) covering every app used by the benchmark.

File	App / Stack	Size
mw-baserow.tar	Baserow	2.86 GB
mw-bigcapital.tar	Bigcapital	2.74 GB
mw-booklore.tar	BookLore	1.45 GB
mw-code-server.tar	code-server	4.60 GB
mw-elabel.tar	e-label	1.86 GB
mw-farmos.tar	farmOS	1.05 GB
mw-grocy.tar	Grocy	273 MB
mw-hrms.tar	HRMS	5.47 GB
mw-mattermost.tar	Mattermost (+Postgres)	1.42 GB
mw-mediacms.tar	MediaCMS	1.64 GB
mw-metabase.tar	Metabase	847 MB
mw-onlyoffice.tar	OnlyOffice Community	9.74 GB
mw-openemr.tar	OpenEMR	2.61 GB
mw-openproject.tar	OpenProject	2.11 GB
mw-opnform.tar	OpnForm	474 MB
mw-owncloud.tar	ownCloud	1.97 GB
mw-photoprism.tar	PhotoPrism	3.56 GB
mw-pretix.tar	Pretix	2.23 GB
mw-recipya.tar	Recipya	593 MB
mw-roundcubemail.tar	Roundcube Mail	1.25 GB
mw-siyuan.tar	SiYuan Notes	2.85 GB
mw-twenty.tar	Twenty CRM	1.97 GB
mw-watcharr.tar	Watcharr	239 MB

Each tar already contains the :latest tag; image names follow the mw-<app>[-<component>] convention so loaders can resolve them deterministically.

Download

from huggingface_hub import snapshot_download

snapshot_download(
    repo_id="Marti844/SaaS-Bench-docker",
    repo_type="dataset",
    local_dir="docker/images",
    allow_patterns=["*.tar"],
)

Or with the CLI:

huggingface-cli download anonymous8722/SaaS-Bench \
    --repo-type dataset --local-dir docker \
    --include "*.tar"

System requirements

• Disk: ≥ 60 GB free for the image archives plus loaded images.
• RAM: ≥ 500 GB recommended if you run the full eval with the default 4-way parallelism — most stacks bundle their own DB / search / document-server, so total memory grows quickly under concurrency.
• Host OS: Linux. Tested on Ubuntu 22.04 and Alibaba Cloud Linux.
• Docker: 24+ with the compose plugin.

Licensing

• This card: Apache 2.0.
• Each bundled Docker image retains the license of its upstream project (e.g. OnlyOffice — AGPLv3, Mattermost — MIT/AGPLv3 dual, OpenEMR — GPLv3, etc.). The images are redistributed for benchmarking convenience only. Verify upstream terms before any non-research use.