Both functions are accessed through Kagi's [[https://help.kagi.com/kagi/api/overview.html][APIs]]. Before a call can be made, some setup should be done on the Kagi website (see below).
- =kagi-fastgpt-shell= :: Opens a shell buffer in a new window where prompts can be typed. This Kagi FastGPT typically returns output based on actual search results. When point is on one of the listed URLs, press =C-c RET= to open it.
- =kagi-fastgpt-prompt= :: Enter a prompt in the minibuffer and show the result in a separate buffer. With a universal prefix (=C-u=), the result is inserted at point. This function can also be used from Lisp code.
The token can be supplied directly as a string, but you could write a lambda to retrieve the token from a more secure location (e.g. with the combination of [[https://passwordstore.org/][pass(1)]] and the password-store package that comes with it).
1. Create a Kagi account if you haven't done so already. An account is free, and comes with 100 trial searches.
2. In [[https://kagi.com/settings?p=billing_api][your account settings]], put a balance for the API part (note that this is a separate balance than the subscription). The recommendation is to start with a one-time charge of $5. A single query ranges from 1 to 5 cents typically, depending on the amount of tokens processed.
3. In [[https://kagi.com/settings?p=api][the API portal]], create an API token. Put the result in ~kagi-api-token~ (or write a function to access it securely).
The code to generate the table of configuration items was inspired by an idea of [[https://xenodium.com/generating-elisp-org-docs/][Álvaro Ramírez]] (a.k.a. xenodium).
Note that, when you invoke the summarizer functionality interactively, you can also temporarily choose a different target language with the universal prefix (=C-u=) on one of the =kagi-summarize-*= commands.
For handling sensitive information and documents, we recommend setting the 'cache' API parameter to False. In this way, the document will "flow through" our infrastructure and will not be retained anywhere after processing.
#+end_quote
In a similar fashion as above, you could define a function that disables caching temporarily (while having it enabled by default).
#+begin_src elisp
(defun my/kagi/sensitive-summary (text)
"Summarize the current TEXT with caching disabled.")
The kagi.el package can be integrated with [[https://github.com/oantolin/embark][Embark]]. Use it to easily summarize, translate or proofread a buffer, region or a URL. In order to be consistent with all keymaps, and to avoid clashes, the functionality is behind the /K/ prefix key. For example, press /K s/ to invoke the summarize functionality.
Since the FastGPT shell inherits from =comint-mode= indirectly, many key bindings are also inherited. Enter the =help= command in the shell to get more info, or run =describe-keymap= on =fastgpt-shell-mode-map=.
One of those bindings is =C-c C-o=, which flushes the last output. However, this binding is used in =org-mode= to open a URL an point. Typical FastGPT results include URLs so one may be tempted to type =C-c C-o= to browse the URL, only to have the output erased.
If you recognize this confusion, you may want to add the following line to your configuration file to shadow the =comint-mode= binding with something more appropriate:
kagi.el comes with some unit tests, written with [[https://github.com/jorgenschaefer/emacs-buttercup/][buttercup]] and can be executed in a controlled [[https://github.com/cask/cask/][Cask]] environment:
1.=git clone https://github.com/cask/cask/=
2.=make -C cask install=
3. Run =cask= in the kagi.el directory to setup the environment.
4. Run the tests with =cask exec buttercup -L .=
There's also a [[https://github.com/casey/just][justfile]] which allows you to execute =just test= to compile the Emacs Lisp source and run the unit tests afterwards in one go.
- Some variables were renamed for consistency. The impact is considered low given the infancy of the package, and API URLs are typically not modified anyway.
- Add boolean variable =kagi-stubbed-responses= to enable stubbed responses, to replace actual API calls. Since calls are metered, it's more economic to use stubbed responses for development / debugging purposes.