DeepSeek + KLINCH

Run the KLINCH agent and chat on DeepSeek models with your own key. Paste the key once, pick a model and start working on any GitHub repository.

Connect DeepSeek in a minute

Open Providers, choose the DeepSeek preset and paste your API key. The base URL is filled in for you. KLINCH loads the model list from the provider, or you can type a model id.

Base URL: https://api.deepseek.com/v1
Protocol: OpenAI Chat Completions with streaming
Key storage: AES-256-GCM, server-side only

Why use DeepSeek for coding

DeepSeek models are known for low prices per token, which makes long agent runs over large repositories affordable.

What you can do with it

Use DeepSeek in agent mode to edit code across files and propose a commit, or in chat mode to ask about @file, @folder or the whole @repo. Switch models between chats whenever a task needs more or less power.

FAQ

Who pays for DeepSeek usage?

You do, directly to DeepSeek, at their prices. KLINCH adds no markup and is free during the beta.

Is my code sent anywhere else?

File contents the agent reads are sent to DeepSeek as part of the prompt. KLINCH itself stores chats and settings, not your repository.