ProductFlow
Self-hosted visual e-commerce workbench for AI copy, image generation sessions, and node-based product workflows.
Build AI product copy, visual posters, and multi-turn image generation workflows on your own server.
ProductFlow is an open-source visual workspace designed for solo merchants and small e-commerce teams. It brings together product data, structured AI text generation, poster templates, and image-to-image workflow nodes on an interactive canvas. You should use it if you want to produce marketing graphics and structured product descriptions using your own infrastructure and API keys.
What it is
The platform provides a node canvas where you arrange product photos, reference images, text generation blocks, and AI image prompts with drag-and-drop connections. It processes background generation requests using a Redis and Dramatiq task queue, supporting branch histories and multi-candidate image comparisons. It also features a mobile-friendly interface with touch controls for editing nodes and reviewing image galleries on the go.
Who it's for
It suits independent online sellers and small creative teams who want total data ownership over their e-commerce visual assets and model configurations. It does not suit larger organisations requiring multi-user access controls, enterprise role management, or direct automated publishing to e-commerce storefronts.
Key features
Interactive Node Canvas
Connect product details, reference photos, structured copy blocks, and image prompt nodes visually on desktop or mobile.
Iterative Image Sessions
Run multi-candidate text-and-image generation rounds with historic base-image selection and side-by-side branch comparisons.
Centralized Asset Gallery
Store and search generated product graphics alongside their parent sessions, prompts, resolution metadata, and models.
Runtime Parameter Control
Configure model vendors, custom templates, image background handling, and fallback options behind a secondary access token.
Background Worker Architecture
Queue asynchronous image and poster jobs with Dramatiq and Redis while tracking state persistence inside PostgreSQL.
Up and running in 3 steps
Our honest verdict
What's great
- Dedicated node canvas specifically built for e-commerce visual workflows.
- Responsive mobile touch controls for viewing and executing workflows on phone screens.
- Protects administrative settings and model credentials behind a secondary security token.
- Includes a mock provider mode to test canvas builds before spending model API credits.
Where it falls short
- Supports only a single admin user with no built-in team role permissions.
- Requires managing a multi-container stack with PostgreSQL, Redis, backend workers, and web services.
- Lacks direct integration for auto-uploading generated images to stores like Shopify or Amazon.
- Requires you to supply and pay for your own commercial AI model API keys.
Pricing & license
ProductFlow is free to use under the MIT license. There are no paid hosted tiers provided by the authors, meaning your only expenses are the infrastructure hosting costs and third-party AI model API usage fees.
Open remote AI-training roles
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Does ProductFlow automatically publish items to online stores?
No, ProductFlow handles asset creation and visual workflow execution locally; it does not include direct store publishing features.
Can multiple team members log in with individual accounts?
No, the current design runs as a single-admin instance protected by a shared access key without multi-tenant permission controls.
Can I test the workbench without active AI API keys?
Yes, ProductFlow defaults to a mock provider mode so you can test node connections and layout creation before connecting paid APIs.
