The Atelier
Build and customize AI characters for your world
Your Characters
The Library
Books apply across bots without being copied into each one. A codex book is shared lore or writing directives (constant entries go into every prompt, keyword entries fire on triggers); a command book is a set of !commands. "My bots" books reach every character you own; admins can publish a book to everyone. A command book set to Just me (Studio and up) follows you instead: any bot you talk to answers your commands, whoever made it. Each character can switch a book — or single commands — off.
Basic Info
Common tags none picked yet
Pick at least one so people can find this character in the Lobby. ✨ Suggest picks them for you.
Personality & Behavior
Advanced Settings
Alternate Greetings
Add multiple opening messages for variety
Book
Codex Entries
The codex is your character's encyclopedia. Add entries for world facts, side characters, locations, and lore — they get injected into the AI's context when relevant keywords come up, so the bot stays on-lore without wasting tokens.
Constant = always active Keyword = trigger words Learn more
Text from your card that replies may repeat as written, like a catchphrase, a song or the notes in a status box. The rest of the card stays protected. Learn more
Custom Commands
Create custom triggers for specific behaviors. Commands from your Library command books are listed under Shared Books below and don't need to be added here.
Storyline beta
A story this character follows, chapter by chapter. It only ever reads the chapter it is in (and a one-line memory of the ones already played), so it can never rush ahead. The platform keeps the clock: it counts the character's replies, counts story days if you ask it to, and every few turns asks the chatter's memory model one yes-or-no question ("has this happened yet?") to decide when to turn the page. Readers see only what you allow.
Shared Books
Library books that reach this character. Switch a whole book off here, or single commands from a command book. A book that reaches all of your characters can be switched on or off for all of them at once. Changes save immediately.
Save the character first to see its shared books.
Test Message
Send one line to this character with your own API key and see the reply, plus the exact prompt it was given and its size. Uses the saved version of the character (save first if you just changed something); nothing is stored in memory.
Show the prompt the model received
Scene Tests
Scenes to run before you publish an edit: an introduction, a boundary being set, a lore trigger, a persona switch, a long scene. Each one is a few lines of history, the persona you speak as and a test line for the character to answer. A run uses your own API key and the saved version of the character, nothing goes into a chat or a memory, and each reply shows beside the last run's with the codex entries that fired.
The same scene won't give the same words twice, so compare how the character answers, not the exact wording.
Save the character first to test scenes.
Import / Export
Import Character Card
Upload a .json card or a SillyTavern character .png (the card inside the image is read, and the picture becomes the avatar)
Bulk Import (drafts)
Drop a whole pile of .json cards, World Info lorebooks, or SillyTavern .png cards. Each becomes a private draft — polish them in passes.
Merge In a Draft
Got a lorebook that imported as its own character? Fold it into this one: its codex moves over, empty fields fill in, and the shell is deleted.
Export Character Card
Download as a TavernCardV2 .json card (saved characters only)
Feedback Inbox
What readers sent you about this character from its profile page and Lobby card, while Reader feedback is on. Only you see it, and a reader's name only when they chose to share it. Mark a note done once it's handled (done notes go after 30 days), or report one that crosses a line to the help desk. Changes save at once.
Save the character first to see its inbox.
Preview
Character preview will appear here...