AI Agent Visual Workbench (containerised edition)

No-code Platforms

A containerised edition of a self-hosted AI agent platform: assemble agents, RAG and workflows by dragging and dropping, try them out in the built-in playground, and serve the finished flow over an HTTP API; what runs is upstream's official Docker image, pinned by digest, and the LLM key is the buyer's own.

Subscribe & download See how to start

SubscriptionUS$49 per month

Last run2026-11-02

In practice

Readiness4 / 5

What you need

Few changes needed4 / 5

Who it is for

Individuals and teams who want an agent workbench standing ready in their own environment: assemble agents and RAG visually, try them out in the playground, and let other systems call the finished flow over an HTTP API.

How to start

$ ./scripts/quickstart/init-env.sh
$ docker compose up -d

Browse the files

Files in the product repository at the pinned commit. A few house files are left out of the zip.

File overview2,000+ paths

▸deploy9

.env.example
.gitignore
README.md
docker-compose.override.yml
docker-compose.yml
observability/
prometheus.yml

▸docker17

.dockerignore
build_and_push.Dockerfile
build_and_push_backend.Dockerfile
build_and_push_ep.Dockerfile
cdk-docker-compose.yml
cdk.Dockerfile
container-cmd-cdk.sh
dev.Dockerfile
dev.docker-compose.yml
dev.start.sh
…

▸docker_example5

Dockerfile
README.md
docker-compose.yml
pre.Dockerfile
pre.docker-compose.yml

▸docs1328

.gitignore
.yarnrc.yml
DEPRECATED_MACOS_SUPPORT_INVESTIGATION.md
README.md
TORCH_MACOS_AMD64_PYTHON313_INVESTIGATION.md
aceconfig.js
agents/
babel.config.js
css/
docs/
…

▸regressions4

1.10.x.yaml
1.11.x.yaml
1.9.x.yaml
README.md

▸scripts102

__init__.py
a11y/
build_component_index.py
check_changes_filter.py
check_deprecated_imports.py
ci/
e2e_deployment_tests/
factory_restart_space.py
gcp/
generate_migration.py
…

▸src390

backend/
bundles/
frontend/
langflow-stepflow/
lfx/
sdk/

▫AGENTS-example.md

▫AGENTS.md

▫ATTRIBUTION.md

▫BUNDLE_API.md

▫CLAUDE.md

▫CODE_OF_CONDUCT.md

▫CONTRIBUTING.md

▫DESIGN.md

▫DEVELOPMENT.md

▫LICENSE

▫LICENSE-99LL.md

▫Makefile

▫Makefile.frontend

▫PRODUCT.md

▫QUICKSTART.md

▫README.md

▫RELEASE.md

▫SECURITY.md

▫SECURITY.upstream.md

▫ci-skip-analysis.md

▫codecov.yml

▫delivery-manifest.txt

▫docker-compose.yml

▫package-lock.json

▫package.json

▫pyproject.toml

▫render.yaml

▫uv.lock

▫wss-generated.config

▫Config and hidden files 116

Latest passing run

Clone → Checkout → Toolchain → Container up

Pre-purchase Disclosure

What you are buying

A containerised, self-hosted AI agent workbench: run the initialisation script to produce the login credentials, then bring up the visual editor behind that login with Docker Compose, drag nodes together into agents, RAG and workflows, and serve the flow over an HTTP API.

A derivative work built on an MIT-licensed open source project, not that project itself. What executes is upstream’s official Docker image, pinned by digest inside the compose file.

Delivery is a frozen snapshot: one zip file, taken from the product repository at revision a29f6e7.

Known limitations

The delivered configuration listens on port 7860 over plain HTTP, with no TLS and no reverse proxy
No model and no API credit is included; until the LLM key you bring yourself has been filled in, running a model node will fail
Data lives in a Docker volume, and docker compose down -v deletes the flows and the database along with it
Once the encryption key in .env is replaced, the LLM credentials previously saved in the interface can no longer be decrypted
The first start pulls the official image from Docker Hub, so it will not come up in an offline environment
The delivered configuration is aimed at single-machine evaluation and internal use: SQLite, one container — it is not a production deployment configuration
Upstream’s front-end pages load fonts from Google Fonts, and the delivered configuration does not change that
Upstream telemetry is on by default and sent to scarf.sh; the delivered configuration has it turned off, and that can be changed back in the compose file
Pinned to the delivered snapshot; it does not follow upstream.

The liability cap, the scope of what is promised and the statutory floor are governed by the Terms of Service.

Vetting record

Everything below was checked against the pinned commit 09ef6b2. Every verdict carries its full reasoning, an evidence link and the checker; failures are published, not hidden — do not take our word for it, verify it yourself.

5 PASS0 WARN0 FAIL

Source ReviewPASS

This product does not execute the vendored source code at runtime: what executes is upstream's official Docker image, pinned by tag plus digest and published by upstream's CI. The files we added (docker-compose.yml / init-env.sh / QUICKSTART) were reviewed line by line by ourselves: they contain no outbound connection and no telemetry code of their own, and the password and encryption key are generated on the spot with openssl; but the telemetry upstream turns on by default was not switched off, and that fact is listed in the known limitations the buyer can see. The upstream source snapshot is both the object of vetting and the basis of the derivative work, and it was not read line by line (the monorepo is enormous); this limit on scope is recorded as it stands.

Script SafetyPASS

The only script that runs on the buyer's path is our own scripts/quickstart/init-env.sh (25 lines, reviewed line by line by ourselves: openssl rand produces two values written into .env, then chmod 600, with no outbound connection and nothing deleted or altered); docker compose mounts no host socket and no host directory (named volumes only). The scripts inside upstream's docker/ and docker_example/ are not on the buyer's quickstart path.

RecencyPASS

At the time of vetting (2026) upstream was not archived and had had new commits that same year (checked with gh api). The source pinned commit 09ef6b2 was the main HEAD on the day of that check; the runtime image is a release published by upstream, and it is not from the same point in time as that source snapshot.

LicensePASS

SPDX:LicenseRef-99LL-Product

Upstream's LICENSE was checked word for word (the evidence link points at the file itself at the pinned commit, not at GitHub's licence field): the standard full text of the MIT licence, Copyright (c) 2024 Langflow — permitting commercial use, modification, derivative works, redistribution and sale, with the obligations being to keep the copyright notice and the full licence text (LICENSE and ATTRIBUTION.md are delivered with the zip). For the product as a whole, the 99LL product licence template is attached: upstream's LICENSE is kept as it stands, and LICENSE-99LL.md is placed alongside it (at the root of the delivered tree, tag v0.1.0-99ll.1) as the sale and liability terms for the parts we added, with spdx recorded as LicenseRef-99LL-Product.

Dependency LicensesPASS

The delivered zip tree vendors no compiled or installed third-party package (neither a python venv nor node_modules is in the tree). The third-party packages used at runtime are inside the official image, packaged by upstream and published on Docker Hub in the name of an MIT project, fetched from there and not distributed with the zip; we did not verify the packages inside that image one by one (recorded as it stands). The files we added have zero third-party dependencies.