> For the complete documentation index, see [llms.txt](https://fictionlab.gitbook.io/fictionlab/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://fictionlab.gitbook.io/fictionlab/getting-started/chatting-page.md). # Chatting in FictionLab Chatting in FictionLab is very easy, but this guide will explain every feature to allow you to maximize your roleplay experience!

### Intro to Chatting The chat is between you and the AI. The AI acts as the other characters while you act as your own character or a pre-defined character, depending on the scenario settings. You can write your responses to the input field at the bottom or choose from two suggested responses if you're unsure what to write on your own. The suggested choices can be disabld from the lightbulb next to the send icon.

*** At the top navigation bar, there are three buttons.

The first one from the left is the **Sharing** button which allows you to share the scenario with others. You can even share a private scenario, allowing people with the link to access it.

The middle button is for **Favorites**. This makes the scenario appear in the "Favorites" section on your user page. Favoriting scenarios allows you to find them easily.

The last button is for **Chat Settings**. Clicking it will open a side menu with several settings, allowing you to customize the chat to your liking, such as changing your model choice, or selecting your response length.

Chat Setting side menu in FictionLab chat

**Maximum number of chats per account:** 500 When you reach the maximum number of chats, FictionLab will display a warning that states that you should close some chats before starting a new one. Older chats will not be automatically deleted by the system. On mobile, you can select multiple chats for deletion. On a browser, chats can only be deleted one by one ***or*** by clearing all chat history under User Profile then Manage Account.

Go to **User Profile** (or Settings) then scroll to the bottom to find **Manage Account** and then **Clear Chat History**

WARNING: * Cleared chats cannot be restored. * Deleted accounts cannot be restored. Account deletion is instant. *** ## LLM Model Selection The LLM model selection allows you to choose between the various AI models that FictionLab offers. Each model has a unique "personality" and writing style, so you can experiment and see what fits your roleplaying style the best. Below are all the current models:

Screenshot of FictionLab AI models — Currently available FictionLab AI models

### Free Models: ### Default The default free AI model of FictionLab. Good creativity and decent memory. Completely unlimited and unfiltered! ### Ophelia Ophelia is an overhauled version of the Aurelius model. It focuses on smaller details and has a slower pace when compared to the default model. Great for slow burn and slice-of-life roleplays. ### Wraithmind This model excels at running darker roleplays. Perfect for horror or violent roleplays where survival isn't always guaranteed. ### Plus Models: ### Ember Ember is a brand-new architecture, fine-tuned-in-house model by FictionLab and built directly on the lessons learned from Glendora. It leans into emotional depth and a slower, more intentional pace, giving tension, romance, and quieter character moments the room they deserve. ### Oracle V3 Based on DeepSeek V3.2, Oracle V3 brings various improvements such as better instruction following, improved memory, and more natural responses. The model is balanced, expressive, and reliable. It delivers masterful storytelling precision with improved narrative consistency and deeper character understanding. ### Chimera Based on GLM V4.6, Chimera is a model that excels in adding variety to the story while focusing on keeping characters true to their definitions. It balances unpredictability with consistency, making it ideal for dynamic stories where characters need to remain grounded despite changing circumstances. ### Quasar Quasar is a fine-tuned DS V4 that delivers response that feel both genuine and spontaneous, as if it knows exactly where you want to take a story. It stays creative within the narrative, generation drama and tension that are always welcome and it excels at slow burn by taking its time to build anticipation. ### Ultra Models: ### Eclipse Eclipse is an Ultra-tier model built on a frontier ensemble tuned for a clean balance of speed, reasoning, and storytelling. It keeps the expressive, adventurous vibe of Oracle while pushing further on coherence, memory, and instruction-following. A strong all-rounder for players who want quick, thoughtful responses without sacrificing depth, well-suited to long arcs, layered characters, and scenes that need both pace and nuance. ### Paragon Paragon harnesses DeepSeek V4 Pro's 1.6T-parameter frontier reasoning to plan scenes, weigh stakes, and follow complex instructions before responding. Tuned by FictionLab for creative writing, it excels at any creative assignment — from tactical roleplays and intricate plotting to sweeping prose and multi-character scenes. Long-context retention is exceptional, and the model maintains a confident, dramatic tone across extended arcs. ### Arcanum Built on the GLM 5.2 architecture, Arcanum is tuned with creativity at its core. The most inventive model in the lineup, it delivers grounded, believable story developments without veering into implausible tangents. It moves naturally between darker, mature themes and lighter moments of genuine humor without losing narrative consistency. Character adherence is one of its strengths: Arcanum expands on established character details, manages multiple independent characters, and preserves hidden secrets while maintaining their individual personalities and motivations. ### Lumina Lumina is an Ultra-tier model built on Xiaomi MiMo V2.5 Pro with reasoning enabled. It is tuned for visceral, descriptive prose. It leans into texture, sensation, and atmosphere; every gesture, scent, and shift in light is rendered in sharp, graphic detail. Even quiet, ordinary moments become charged and sensational, making it ideal for writers who want their scenes to truly come alive on the page. ### Wayfarer Wayfarer is an Ultra-tier model fully ready for your role-play. It takes the lead and moves the story forward adding surprising events, twists and new lore so things keep happening even if you type short replies. That makes it a great pick if you want the AI to do the heavy living. It writes in a rich, detailed style and gives each character their own voice, from the main cast to background NPCs, and it handles busy group scenes as well. *** ## Chat Response Settings * **Response Length -** Allows you to specify your preferred AI response length. The options are Long, Medium and Short. The default response length is Medium. * **Stream Text Response** - Enables a typing animation where the generated text is shown to the user in real time. With this setting enabled, the response will only appear when the response is completed. * **Show Response Suggestions** - Enables or disables the two pre-generated choices after receiving a response from the bot. *** ## User Character The User-Character is the character you, the User, will play as during the scenario. You can access the User-Character menu from the chat settings under "Edit User Character" and edit it as the story goes on.

User Character Menu. Old image that still shows 15 slots when paid now have 200 slots.

Hovering over the User-Character cards brings up three dots. Clicking on these dots allows you to either delete or edit the character card. User-Character cards should be written in the same way as Story Character cards are. Describe the looks and personality in the description. More detail usually gives better results. Free users are limited to **10** User-Character slots. FictionLab+ users have **200** User-Character slots. If you do not select a Playable Story Character or a User-Character, the following warning will appear when you try to chat:

Warning message when there is no selected User Character

Just click your chat settings to select a User-Character. *However*, this option will not be available for scenarios that use Playable Story Characters only. In which case, you need to click Start Scenario again. {% hint style="warning" %} If you select a Playable Story Character, do not exit that character because you cannot select it again. Playable Story Characters are Select at Start Only. {% endhint %} *** ## Custom LLM Instructions

Custom LLM Instructions in Chat Settings

Custom LLM instructions are almost identical to custom instructions in the scenario creation page. You can use it to guide or instruct the AI to act how you want. Instructions should be clear and concise for the best results. The limit for Custom LLM Instructions for free users is 300 characters, and 1000 for FictionLab+ users. Custom LLM instructions should be used to nudge the AI in a certain direction. When instructing the AI to *avoid doing something*, it is ideal to *give it directions as to what it should do instead* as shown in the example below. Examples: * Narrate in third person perspective. * Refrain from repeating {{user}}'s dialogue or actions. Instead, move the story forward by reacting to {{user}}'s message. *** ## Narration Voice Settings

The narration voice setting allows you to change the voice that reads messages when using the narration feature. You can activate narration by clicking on the speaker icon while hovering above the AI's responses. Enabling the Auto-Play Voice Narration toggle will automatically start narrating the message after it has finished generating. *** ## Translation Language Setting The translation language setting allows you to choose between supported languages for the translation feature.

After selecting a language, hover above the message you want to translate and click the translate icon at the bottom of the chat message. It should generate a new message which is a translation of the original chat message.

*** ## Chat Font Size The chat font size allows users to customize the font size of the messages in this specific chat window. It ranges from 10% all the way to 200%. *** ## Response Typing Speed The response typing speed allows users to customize the text typing animation speed. The typing speed is limited by the generation speed of the AI. At the fastest setting, it will write out the text at the same time as it is being generated. Larger speeds can lead to pauses between typing. It ranges from 10% all the way to 200%. *** ## Chat Color Customization The color customization settings allow you to customize the color palette of the chat exactly to your liking. You can select any normal hex color using the color palette.

*** ## Chat Background Style Chat background styles allow you to customize the chat window to fit your roleplay better and make your roleplay more immersive. Free users have access to 7 backgrounds, while FictionLab+ users have access to over 40 different backgrounds. FictionLab+ users can also upload custom backgrounds from their device.

## Locations

Locations are a feature to show you where you are and display a creator added background. The locations will have a Flourish at the top of the screen letting you know where you are. Creators will have to upload an image in 1280x720 and enable behind the scenes.

Free, Pro and Ultra users will be able to choose between their own background and the creator background. Ultra users still have the option of the Dynamic Backgrounds, or AI Scene generation. {% hint style="info" %} If you do not wish to see the Flourish of the location at the top you can disable in your chat settings. {% endhint %}

Here is an example of a generated background:

{% hint style="info" %} If a location is set to Hidden, the image will not display. {% endhint %} *** ## Chat Formatting FictionLab uses a variety of different symbols to indicate different behaviors. For example: * `*single asterisks*` italicize text * `**double asterisks**` make text bold * `"quotes"` wrap the text in double quotation marks and often indicate dialogue * `` `single backticks` `` makes inline code. * ` ```triple backticks``` ` make a code block For your own messages, normal text is usually action and dialogue are in quotes. Italics can be used for emphasis, internal thoughts or language translation. The AI is very lenient with formatting, so you can utilize any formatting style you prefer. *** ## Message Menu & Editing

Hovering over the top of a message show three vertical dots. Clicking on these dots displays menu. This message menu allows you to copy, edit, branch, rewind, and delete messages. You can also regenerate the message using this menu. #### Chat Editing Editing the message allows you to change parts of the message to your liking. It can be used to change events that have happened during the roleplay and to drive the roleplay in a different direction. Use the previously mentioned formatting for best results when editing.

Messages can also be reported if you find them to break our TOS and guidelines. Our staff team will then review the message and take action if needed. *** ### Chat Branching Chats can be branched from the message menu by selecting "**Branch From Here**". This will create an identical chat session up to that point, so you can experiment with taking different paths in the roleplay.

### Timelines & Saved Chats You have the ability to swap your Timeline (from the Chat Branch) at any time. Simply select which timeline you want to write with and the scene updates at the flick of a button.

{% hint style="info" %} Chats are capable of being organized to your liking. You can Pin your favorite chats, select View All to Search through your chats, Rename or Archive! {% endhint %}

#### Message Regenerating Messages can be regenerated if you're not happy with the results. You can press the reload symbol that appears at the bottom of the message while hovering to regenerate the message. After regenerating, you can switch between the response versions using the arrow symbols at the bottom of the message box.

## Renaming Chats For Desktop: You can rename the current chat session by hovering over the chat button in the side menu.

For Mobile: You can access the same rename menu by pressing on the chat button in the Chats page and selecting "Rename Scenario" next to the 3 Dots of your scenario name. *** ## Exporting Chats Users can export the chat from FictionLab by opening the chat settings menu and scrolling all the way down. There, you can select "Export as PDF" or "Export as JSON". *** ## Slash Commands FictionLab offers a slash command system to make using certain features easier and quicker. You can see all of the available slash commands by typing "/" in the chat input bar.

* **/help** : Shows a list of the current slash commands. * **/continue** : Ask the AI to continue from the last response. * **/favorite** : Favorite the current scenario. * **/improve TEXT** : Rewrite your draft message to better match your User-Character. * **/memory-cards N** : Manually create memory cards from the last N messages for example /memory-cards 10 * **/model NAME** : Change the LLM currently used for example /model glendora * **/narrate** : Narrate the latest message. This activates the voice narration. * **/ooc TEXT** : Out-of-character instruction. This is a very useful command that can create anything from summaries to directing the LLM. * **/pov NAME** : Generate a response from the perspective of a character. * **/recap N/ALL** : Generates a recap of N past messages or ALL past messages. This can be used in combination with /memory-cards to create manual memory cards by copying the output of /recap. * **/theme NAME** : Change the chat theme for example /theme black \\ *** *** ## Disabling Auto-Keyboard Popup You can change the keyboard behavior when the response finishes generating through the settings menu.

*** ## Change Default Chat Settings Users can adjust the default chat settings that are applied when they start a new chat from the settings menu. The settings are identical to the ones shown in the chat settings side menu.