Zammad helpdesk AI features with a self-hosted LLM (Ollama / OpenAI-compatible)
Zammad 7 adds ticket summaries, AI-assisted replies and auto titles; a hosting provider's docs show connecting it to Ollama or any OpenAI-compatible self-hosted model via Settings > AI > Provider.
2-10 people11-50 peopleCustomer support$3k+Sensitivity: High (client/confidential)
Hardware
| Item | Approx. price (as of 2026-10-16) |
|---|---|
| GPU server for the LLM ('GPUs with typically >= 24 GB VRAM' per source) | Not stated in source |
Source prices are as stated by the source. Retailer offers: prices checked 2026-10-16, may change. Some links may be affiliate links; we may earn a commission at no extra cost to you. This never changes what we list.
Software stack
- Zammad 7
- Ollama or any OpenAI-compatible server
Models used
- Any self-hosted model; provider also lists GPT-OSS 120b, Mistral Small 3.2 24b, Llama 3.3 70b
Good at
- Built-in ticket summaries and AI-assisted replies
- Only needs API URL, token and model name
Weak at
- LLM API must be reachable over the internet (with authentication) for hosted Zammad
- Source warns running an LLM is resource- and cost-intensive
Rough total cost
Not stated in source.
Getting started
- Run a self-hosted model behind an authenticated OpenAI-compatible or Ollama API.
- In Zammad: Settings > AI > Provider.
- Enter API URL, token and model name.
- Enable summaries / AI-assisted replies for agents.
Source & credit
Zammad - KI-Features & LLM-Integration (German) — server.camp Docs
Summarized from the source; quotes are the source author's words. Check the original before buying.