Assistant Settings

Slek includes a built-in assistant that can browse and modify your training data through conversation. The assistant has several parameters that significantly affect both the quality of results and the cost of each request. That is why it is up to you to choose the right balance for your needs.

In addition to the built-in assistant, Slek allows you to use an external AI via an MCP server (see MCP Server below).

Provider — Various AI providers provide different performance/price tradeoffs.

Quality (Claude only) — Three levels (Affordable, Standard, Premium) that select a more or less capable model. Higher quality means better results but a higher cost per token.

Effort (Claude only) — How much effort the model puts into each response. More effort produces better but costlier responses.

Thinking / Reasoning (some models) — When enabled, the model reasons step by step before answering. Useful for complex tasks (e.g. season planning) but more expensive.

Response style — Precise, Balanced or Creative. Controls how varied and creative the assistant's writing is.

Running an assistant has a real cost: every request consumes computing resources billed by the AI provider. To keep things fair, Slek charges for the built-in assistant on a per-use basis — each user pays in proportion to their actual consumption. The costs shown here only apply to the built-in assistant.

An MCP server is an application that lets an AI connect to read data and perform operations. Using the Slek MCP server is free (it only requires a Premium subscription). When you use MCP, the AI cost is handled by your own provider (Claude, Mistral, etc.) and is typically cheaper, because major AI companies invest heavily in competitive pricing to attract users.

Slek recommends Claude and Mistral for the best MCP experience. If integration with another provider is important to you, don't hesitate to contact support.