Project Alterra | test

🔥 I run some experiments with AI front memory prompting. Tell me in comments if you notice any changes. ⚠ This scenario is premium only. Free models are too weak to pick up new formating. ⚠ Because of script that removes repetitiveness, you can't go past ~24k context. It's some weird bug with context devs should fix. ❕ Project Alterra heavily rely on "Dungeon Extension" created by discord user named Claudia. https://discord.com/channels/903327676884979802/1417245468542636062/1417245468542636062 ℹ️ This scenario features system designed to override default AI dialogue behavior and enforce custom dialogue narration. Scripts replace any default output with error message, forcing the AI to abandon its usual formatting and adopt the new structure. 🔷Scripts this scenario has: Default dialogue erasure script: Automatically deletes any default-style dialogue and replaces it with error message. Banned words: Story card in which you can write any word or/and sentence to make it automatically get deleted in next outputs. Replace words: Similar as banned words, but replaces one word/sentence onto the other of your choosing. Judgebot: interrupts the story every N turns to judge a recent moment with humorous commentary and point scoring. Can be adjusted in "[CONFIG] Judgebot" story card or turned off by setting "Judgebot Frequency" to 9999 Paragraph fix script: Makes story more readable. Narrative Guidance Overhaul: https://discord.com/channels/903327676884979802/1380157986709962772 ♦️ Why this formating is better: It reduces tokens consumption by removing repetitive nonsense that usually cling to dialogues. Thus, saving you tokens. Tokens = AI memory. Now your story can be focused on more interesting things than He/she says while tapping fingers, while smirking. while dripping with amusement, while smiling even wider, while whispering/barking non-stop and etc. all that nonsense. ✨ User tip: Try out "Qw" command. Write "Qw Current weather"