Select a JSON or PNG file to be imported

[Elo] [8/8/25] 4.5 has kicked the bucket. Sorry, friends. GPT5 is good to go and like some COT models the temp is set and streaming is turned off. This is not my choice, but an OAI level restriction. Price reflection might not be accurate at this time. More updates soon! <3

[Elo] [6/11/25] Model update has dropped! We've expanded on the available o1 and o3 models, as well as giving the list the usual tidying up to add any new model snapshots and remove non-fucntional ones. Quick reminder, o3 is only available for greenlit organizations, standard accounts can not access this model.

[Elo] [6/01/25] Patch 2.3.3 is now live! The updates Turtle and I have been working on the last two months are finished. If you encounter any bugs, as per usual, please let us know via discord or email.

[Elo] [5/29/25] Finished the Prompt/Opinion panel overhaul. Added inner thoughts and a thread recap option! I might be tweaking the prompt side of these a little bit so they function smoother as time goes on.

[Elo] [5/28/25] The prompt window has been renamed the options window. Banned words and a disable OOC feature have been added along-side an optional safe word function. Debugging and some minor visual tweaks were made. New buttons coming next week!

[Elo] [4/20/24] Began the overhaul on the opinion panel!

[Elo] [2/28/25] GPT 4.5 is live! Be careful! It’s expensive!

[Elo] [2/16/25] The 2.1.2 update makes it so that users can now use o‑series models without any fancy tricks. The last patch did not properly address every pitfall to being able to interact with this model. This has been fixed ♥

[Elo] [1/24/25] Phew! 2.1.1 is good to go! o1, o1 mini and preview now work as intended. While we were able to ping these models before, messages did not properly send and receive. This has been fixed.

[Elo] [12/15/24] 2.0 is live! You might notice an additional button on the prompt panel — this is an alternative greeting button. Store infinite alt introductions locally, change them out at the click of a button, or choose one of our supplied alternative greetings to go on a variety of different AI adventures with your characters. For now, the alternative introductions offered by us are blank and will be coming over the next couple of days while we ensure that this feature is working smoothly for user‑created greetings. Have fun!

[Elo] [12/15/24] December patch is slowly being rolled out. All dead links have been updated. If you notice any issues, as always, please report them in Discord! Thanks!

[Elo] [10/30/24] We’ve continued dropping some new characters and bots over the course of October! Feel free to join us on YouTube or Discord to check out even more content.

[Elo] [10/5/24] Another content‑jam drop has landed with 1.9.8! We’re introducing two more functional bots at the bottom of the page: a coding assistant and a bot that replicates the COT or o‑1 strawberry model from OpenAI.

[Elo] [10/3/24] It’s October! And that means the content jam has started with 1.9.7 introducing two new characters! We’ll be pushing out a number of characters and QoL improvements over this month. We all hope you enjoy! ♥

[Elo] [9/24/24] 1.9.6 — Universal Patch has been implemented. Old alpha versions now utilise beta 1.9.6 code to help resolve edge‑case bugs that occurred between old and new versions communicating. This does not change your chats, characters, keys, or any aspect of your data. Thanks!

[Elo] [9/20/24] Right! SO! 1.9.4 was buggy, that’s been fixed with 1.9.5 — I’m really sorry this didn’t get out sooner! Model rights now flag properly and chatgpt‑latest is working again.

[Elo] [9/13/24] Well it’s about darn time 1.9.4 got here. It’s taken a bit of work to get o1-preview and o1-mini up and running alongside chatgpt‑4o‑latest! If you encounter any errors, make sure to report them on Discord so they can be addressed right away!

[Elo] [8/6/24] Say hello to 1.9.3 and GPT‑4 Long Output! This model is expensive as all get‑out but promises longer outputs. This is an alpha model not everyone has access to currently. I’ve changed the lore and summarisation prompts in hopes of preserving the original language of the conversation, something our international users have been asking for. At the moment it is still translating to English; I’ll continue to search for fixes. If you find any problems, I’m happy to help!

[Elo] [7/18/24] Hey hey! GPT‑4‑o Mini has been added — and some small bug‑fixes have come with it. A recent bug regarding model context (thanks Jenz!) has been solved and some tooltips updated. See you guys in early Aug!

[Elo] [7/8/24] Great news everyone! We finally have different model options for handling memory and lore now — yay! I’ve also cleaned up the connections for all GPT models. More updates are coming early Aug! If you run into any issues from this update or any other, please remember to report them on Discord!

[Elo] [5/14/24] GPT‑o added and functioning in the model list. Costs are not 100 % accurate with this model yet. GPT‑instruct removed due to no longer being in service.

[Elo] [3/7/24] Jess and Tarek (Nyar) character cards have been added to both TTI and cAI. More characters and a potential art overhaul will be coming soon, as well as “in development” tags for certain characters that may be subject to upcoming changes when first released. Anyway! See you guys soon!

[Elo] [2/22/24] You’re probably thinking things look just a little different — and you’re right! Thanks to our new dev Turtle Guy we’ve got a much more responsive prompt screen and a new news tab where you can keep an eye on the latest in development. We’ve obviously moved the prompt button! It now sits just beneath your chat threads in the bottom‑left of your screen. Thanks for checking in, we’ll see you soon for a new update. ♥

[Elo] [2/22/24] Turtle Guy’s news panel and prompt‑screen fixes were blended with Bimble’s code decompression for more reasonable development and some minor hot‑fixes to styling.

[Elo] [1/12/24] Lorebooks now work correctly. I moved the proxy server that handles lorebooks over to our main server after rewriting its code. Have fun!