1 files changed, 51 insertions, 0 deletions
diff --git a/modules/ai-config.el b/modules/ai-config.el
new file mode 100644
index 00000000..14380a2b
--- /dev/null
+++ b/modules/ai-config.el
@@ -0,0 +1,51 @@
+;;; ai-config.el --- Configuration for AI Integrations -*- lexical-binding: t; -*-
+
+;;; Commentary:
+
+;; There are several workflows available. Here are the ones I use most.
+
+;; - Launch GPTel and chat with the AI in a separate buffer.
+;;   Chatting is fine, but it can mean cutting and pasting code back and forth
+;;   between buffers.
+
+;; - Select a region and launch GPTel.
+;;   The region is automatically inserted into the buffer making it easy to
+;;   simply ask a question after it's read the code.
+
+;; Note that you can save a file, then turn on gptel-mode to resume your
+;; conversation.
+
+;; Remember that sending the message requires C-<return>.
+
+;;; Code:
+
+;; ----------------------------------- GPTel -----------------------------------
+;; integration with ChatGPT and other large language models.
+
+(use-package gptel
+  :defer .5
+  :bind
+  ("C-h G" . gptel)
+  (:map gptel-mode-map
+        ("C-<return>" . gptel-send))
+  :custom
+  ;;  (gptel-model "gpt-3.5-turbo-16k") ;; next best alternative
+  (gptel-model "gpt-4")
+  (gptel-default-mode 'org-mode)
+  :config
+  (setq gptel-directives
+   '((default
+	 . "You are a large language model living in Emacs and a careful and
+  knowledgeable emacs-lisp programmer. Respond accurately and concisely.")
+	(programming
+	 . "You are a large language model and a careful programmer. Provide code
+  and only code as output without any additional text, prompt or note.")
+	(writing
+	 . "You are a large language model and a writing assistant. Respond concisely.")
+	(chat
+	 . "You are a large language model and a conversation partner. Respond concisely.")))
+
+  (setq gptel-api-key (auth-source-pick-first-password :host "api.openai.com")))
+
+(provide 'ai-config)
+;;; ai-config.el ends here.