classiflow-platform

Complex classifications made easy. Open-source, AI-assisted, and user-friendly. Built for ISCO, ISIC, COICOP, and beyond.

classiflow-platform is the front door for the classiflow ecosystem: documentation, local development orchestration, and integration guidance for running classiflow with the optional taxomind AI service.

AI is optional. classiflow runs as a full labeling platform without taxomind.

Components

• classiflow: labeling UI + database + workflows
• taxomind: optional AI classification service (HTTP API)

Links

• Website: https://rowsquared.com/classiflow/
• classiflow repo: https://github.com/rowsquared/classifAI
• taxomind repo: https://github.com/rowsquared/taxomind

Why classiflow

Faster & Cheaper

Get instant preliminary results from AI, then let your team verify and correct the suggestions. Time and cost drop as the model learns from corrections.

Improved Quality

Review and correct AI suggestions to continuously improve quality. Use review and double-annotation workflows to create gold standards.

Simple Process Management

Track progress, assign work, and manage workflows in one place. Custom filters, flexible assignments, and built-in commenting keep teams aligned.

Solid AI Integration

AI predictions can be wrong. classiflow (via taxomind) keeps humans in control, so teams review every prediction and teach the model over time.

Architecture summary

+-----------+        +-------------------+        +------------------------+
| Web users | <----> | classiflow UI/API  | -----> | taxomind API (optional) |
+-----------+        +-------------------+        +------------------------+
                           |
                           v
                      +---------+
                      | Postgres|
                      +---------+

• classiflow is the system of record (users, projects, annotations, audits).
• taxomind provides optional AI label suggestions over an HTTP API.
• classiflow authenticates to taxomind with a Bearer token.

Quickstart (Docker Compose)

Docker images are not published for this project. The provided dev/docker-compose.yml builds images locally from source. If Docker or Git are missing, install them first.

Prerequisites

• Docker + Docker Compose
• Git

The compose file expects classiflow-platform, classiflow, and taxomind to be cloned side by side. If they are not, local builds will fail.

1. Clone all three repos side by side:

   mkdir -p classiflow-workspace
   cd classiflow-workspace
   git clone https://github.com/rowsquared/classiflow-platform.git
   git clone https://github.com/rowsquared/classifAI.git
   git clone https://github.com/rowsquared/taxomind.git

2. Enter the platform repo:

   cd classiflow-platform

3. Create your local env file:

   cp examples/.env.example .env

4. Start classiflow (default, no AI):

   docker compose -f dev/docker-compose.yml up -d

5. Open classiflow:
   • http://localhost:3000

How to use this platform

• Path A: Run classiflow only (no AI). Use the default compose command (no profile). Leave AI_LABELING_API_URL and AI_LABELING_API_KEY unset or blank. See the classiflow README.
• Path B: Run classiflow + taxomind (AI suggestions). Set AI_LABELING_API_URL=http://taxomind:8000 and AI_LABELING_API_KEY=change-me, then run docker compose -f dev/docker-compose.yml --profile ai up -d. See the classiflow README and taxomind README.
• Path C: Work directly with taxomind (advanced / technical). If you are integrating taxomind into another system or developing models, work in the taxomind repo directly. See the taxomind README.

Manual setup (advanced)

• Run classiflow from its repository (UI + API + DB).
• Run taxomind from its repository if you want AI suggestions.
• Point classiflow to taxomind with AI_LABELING_API_URL and AI_LABELING_API_KEY.

Configuration overview

• .env at the repo root drives dev/docker-compose.yml.
• Common classiflow settings: NEXTAUTH_SECRET, DEFAULT_ADMIN_EMAIL, DEFAULT_ADMIN_PASSWORD.
• AI integration settings: AI_LABELING_API_URL, AI_LABELING_API_KEY.
• For full configuration options, see the classiflow and taxomind docs.

Documentation

• docs/usage.md - responsibilities and interaction overview
• docs/architecture.md - components, ports, and integration boundaries
• docs/local-dev.md - local dev workflow and troubleshooting
• docs/deployment.md - production patterns and scaling
• docs/integration.md - classiflow to taxomind configuration

Project status and maintenance

This project is published as an open-source reference implementation and working tool.

We welcome bug reports, documentation fixes, and general feedback via GitHub issues. However, please note that maintenance is intentionally kept at a low-touch level.

New feature requests, refactors, or roadmap-driven changes are not guaranteed to be addressed unless they align with our own priorities or ongoing work.

The software is provided as-is, without any commitment to continuous maintenance, backward compatibility, or support timelines.

Support

Open an issue in the relevant repository:

• UI/workflows: classiflow repo
• AI service: taxomind repo
• Stack/docs: this repository (classiflow-platform)

License

MIT. See LICENSE.