AI assistants have found a new home in instant messaging apps. We previously talked about LuzIA, which has excellent conversation skills, and TranscribeMe is ideal for turning WhatsApp voice messages into text. Now Cami joins the scene, and besides incorporating "personality modes", it also features an image generation system based on Stable Diffusion.
If we think about it for a moment, it makes more sense than it seems: Instant messengers like WhatsApp or Telegram and the new AI models have a very similar dynamic. In fact, I think it's only a matter of time before Facebook/Meta incorporates a native solution in WhatsApp, but after trying some external bots, the truth is they don't work that badly. Today it's Cami's turn, which comes to our devices with a couple of tricks up its sleeve.
Cami: Chatbot, Transcriptions, and Stable Diffusion in Your WhatsApp
Cami requires no additional software: It installs like an extra contact in our messengers and uses a series of six basic commands that we can learn with the /help option.
- /personality: This allows the chatbot to adopt the personality of a fictional character or a famous person. I admit I haven't deeply evaluated this function, but I imagine it can lead to very colorful conversations.
- /language: This command helps us define the language. If for some reason Cami responds in English, a change to /language español should fix that detail.
- /imagine: This is without a doubt the most interesting command, as it enables access to the Stable Diffusion model to generate images in WhatsApp. Cami recommends short prompts, but we can use much broader descriptions without problems (with the exception of some forbidden word). It works better in English.
- /reset: This command does exactly what it promises: It deletes all messages and restarts the conversation as if it never happened.
- /model: This allows us to change the chatbot model. The only variable available for the free mode is ChatGPT, but premium users have the possibility to chat with GPT-4.
- /subscribe: This takes us to the purchase page for a subscription. The free mode is limited to 100 messages every three hours and three generated images in the same interval. A premium account raises those limits to 100 messages and 15 images every three hours, in addition to activating GPT-4 (just 10 messages).
One last detail to mention is that Cami correctly interprets the audio messages we send it; however, it also says it cannot transcribe other voice notes... which is false. Maybe its developers want to polish that a bit, but overall, Cami is on par with other similar assistants.
Official site: Click here
