¶ AI features
¶ Overview
The AI features page is where you turn Splynx's AI integration on and connect it to a language model (LLM).
From here you enable individual AI-powered features and choose which provider processes the requests. You can reach it from Config → Main → AI:
Before you configure AI features, make sure that you have the credentials (such as an API key) for any third-party provider you plan to use, if you are not using the managed Splynx AI provider.
¶ AI Features
The AI Features section controls which AI capabilities are active.
- Enable AI features — the master switch for all AI functionality. The other options on this page work only while this toggle is on. Use the
resetbutton next to it to return the AI settings to their defaults. - Enable AI ticket summary — generates AI-powered summaries for tickets when new messages arrive.
- Enable AI text actions — lets agents polish, shorten, or elaborate reply drafts with AI assistance.
After enabling AI features for the first time, the required AI components are prepared in the background. Give the system a moment to finish before the dependent features become available.
¶ LLM Provider
In the LLM Provider section you choose which language model processes AI requests. Open the Provider drop-down and select one of the available options:
- Splynx AI (managed) — the default, hosted and managed by Splynx. No extra credentials are required.
- OpenAI — connect any provider that works over the OpenAI API protocol. You can point Splynx at any OpenAI-compatible LLM provider by entering that provider's API key and its Base URL (OpenAI-compatible).
- Google Gemini — use your own Google Gemini account.
When you select OpenAI or Google Gemini, additional fields appear so you can enter the credentials for that provider:
- API key — the API key for the selected provider. The value is stored encrypted, so after saving the field hides it and shows a "This field has a hidden encrypted value" notice instead of the key.
- Base URL (OpenAI-compatible) — the base URL of your OpenAI-compatible provider, including the version path (for example,
https://api.openai.com/v1). - Model — choose the model that processes your requests. If your API key has access to the provider's model list, select a model from the drop-down. Some keys grant chat access (all Splynx needs to work) but not access to the model list — in that case the drop-down is empty and you type the model name manually, for example
openai-gpt-oss-120b, or any other model you know the key can use.
After entering the name, click Test connection to confirm the key works with that model.
Click Save to apply your settings.
¶ Important notes
Selecting a third-party provider (OpenAI or Google Gemini) means ticket and message content is sent to that external service for processing. Use the managed Splynx AI (managed) provider if you prefer to keep processing within Splynx's managed service.
The Enable AI features master toggle gates every other AI feature in Splynx. If AI-powered features such as ticket summaries or text actions are not working, check that this toggle is on first.