Products Overview
Use this page to see what’s available today, what’s hosted by us, and what’s coming next. Open Ticket AI is the flagship on-prem product; models and APIs are optional add-ons.
At a glance
| Product | What it is | Status | Links |
|---|---|---|---|
| Open Ticket AI (On-Prem/Main Product) | Local, open-source ticket classifier (queues & priority) integrated via pipelines/adapters. | ✅ Available | Overview |
| Hosted Prediction API (German) | HTTP API to classify queue & priority using our public German base model (hosted by us). | ✅ Free for now | API Docs |
| Public Base Models (German) | Base models for queue/priority published on Hugging Face for users without their own data. | ✅ Available | See links in API Docs |
| Synthetic Data Generator | Python tool to create multilingual synthetic ticket datasets; planned LGPL. | ✅ Available | Generator |
| Ticket Datasets (v5, v4, v3) | Synthetic datasets made with our generator (EN/DE focus in v5/v4; more langs in v3). | ✅ Available | Dataset |
| English Prediction Model | Base model for EN queue/priority. | 🚧 Coming soon | (will be added here) |
| Additional Languages & Attributes | Models for other languages; predictions for tags, assignee; optional first-answer. | 🧭 Exploring | (roadmap) |
| Web UI for Data Generator | Browser UI on top of the generator for non-technical users. | 🧭 Exploring | (roadmap) |
Pricing note: The hosted German Prediction API is currently free. If demand drives infra costs too high, we may introduce rate limits or pricing. On-prem Open Ticket AI remains open-source and local.
Open Ticket AI (On-Prem/Main Product)
- Runs locally; integrates with Znuny/OTRS/OTOBO via adapters.
- Classifies Queue & Priority on inbound tickets; extensible pipeline architecture.
- Pairs well with our Synthetic Data Generator for cold-start or class balancing.
Learn more:Overview
Hosted Prediction API & Public Base Models (German)
- For teams without their own data where the base queues/priorities fit reasonably well.
- Use the German model via our hosted API (free for now).
- Models are public on Hugging Face; you can also self-host or fine-tune.
Start here: Prediction API
Synthetic Data Generator
- Python tool to create realistic, labeled ticket datasets (subject, body, queue, priority, type, tags, language, first answer).
- Planned LGPL release; email for access or modifications: sales@softoft.de.
Details: Synthetic Data Generation
Ticket Datasets
- Multiple versions available:
- v5 / v4: EN & DE, largest and most diverse.
- v3: more languages (e.g., FR/ES/PT), smaller.
- Ideal for bootstrapping, benchmarking, and multilingual experiments.
Browse: Multilingual Customer Support Tickets
Roadmap
- English base model for queue/priority (hosted & downloadable).
- Optional models for other languages.
- Additional attributes: tags, assignee, and first-answer generation.
- Early prototype of a web interface for the data generator.
FAQ
Is the API part of Open Ticket AI? No. Open Ticket AI runs locally. The Prediction API is a separate, hosted service that uses our public models.
Can I bring my own taxonomy? Yes. Train locally with your data, or ask us to generate synthetic data that mirrors your queues/priorities.
Support & Services? We offer support subscriptions and custom integrations. Contact sales@softoft.de.