> For the complete documentation index, see [llms.txt](https://powerfan-1.gitbook.io/redchat.app/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://powerfan-1.gitbook.io/redchat.app/guides/quick-creation.md). # Quick Creation Facing a blank screen for your very first character can feel a little overwhelming. So many fields, and no clear sense of what to fill in or what actually matters. That's okay — **everyone feels that way at first.** This guide will take you by the hand and walk you all the way from that empty screen to **publishing your first living, breathing character.** First, the good news. Only a handful of fields truly matter. The rest is seasoning that adds flavor. And **you can always go back and fix anything later** — so don't freeze up trying to make it perfect. The fastest path is to fill everything in once, start to finish. > **By the end of this guide:** you'll be able to build and publish a character from scratch. When you start getting curious about deeper "writing craft," continue on to [Advanced Creation](/redchat.app/guides/advanced-creation.md). This document covers the **fundamentals**; that one covers the **advanced techniques.** Throughout this guide, we'll build **one example character** together. Her name is **Hana** — a barista at a small neighborhood café. Watching how Hana takes shape at each step will make filling in your own character far easier. *** ## Before We Start: Just One Big Idea The fields in character creation split into **two kinds**. Get this one distinction into your head before you begin, and everything else falls into place. | Kind | Which fields | Who they're for | | ---------------------------- | ---------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------- | | **Fields people see** | Short Bio, Character Bio (long) | For **people browsing.** Has **no effect** on AI behavior | | **Fields that drive the AI** | Description, First Message (and, in advanced use, Scenario · Example Dialogs · Lorebook) | For the **AI.** These determine how the character actually talks and behaves | > **Why this matters:** this is the single biggest point of confusion for beginners. **No matter how much personality you pour into Short Bio and Character Bio, the AI never reads it.** Those fields are the "billboard" that dresses up the explore screen and profile page. A character's real personality has to go in the **Description** to actually work. Remember just this one thing and the quality of your first character changes dramatically. Alright — let's go through it in the order you'll actually build it. *** ## 1. Profile Image — Required **This is your character's face.** It's the avatar that shows up first everywhere: the explore screen, the chat room, the profile page. * It's **required.** A character needs a face. * **Why it matters:** people see images before they read words. A good profile image creates that "let me tap this" first impression. > **Tip:** pick an image whose expression and mood match the character's personality. If you use a beaming smile for a calm, reserved character, browsers will get mixed signals. Images that are close to square with a clear face come out cleanest. *Hana (example):* I picked a warm-feeling barista image — apron on, a slight smile from behind the café counter. *** ## 2. Background Image — Optional **This is the backdrop of the chat room** — the scenery behind the conversation when you chat with the character. * It's **optional.** The character works fine without it. * **Why it matters:** it adds atmosphere. A café interior for a café character, a forest or castle for a fantasy character — a fitting background brings the immersion to life. > **Tip:** it's perfectly fine to skip this on your first build. Add a background you love later, whenever you find one — because **you can always edit it.** *Hana (example):* I set the background to a photo of a cozy café interior with warm lighting. *** ## 3. Name — Required **This is your character's name** — the one you'll call them by in chat. * It's **required.** * **Why it matters:** the name is the character's identity. An easy-to-say, memorable name is best. > **Tip:** avoid names that are too long or complicated. A short, crisp name reads well in chat and in the explore list alike. *Hana (example):* `Hana` *** ## 4. Short Bio — Required **This is the one-line intro shown on the explore screen and profile page.** It's the **hook** that makes people go "ooh, this looks fun" and click. * It's **required.** * **This is a field people see. It has no effect on AI behavior.** Think of it as ad copy meant to draw browsers in. > **Heads up (remember this):** even if you write personality notes in the Short Bio — "she's shy, sweet, and loves coffee" — **the AI does not read it.** Personality has to go in field 6, the **Description**, to work. The Short Bio exists purely **to earn a human's click.** > **Tip:** don't explain — make them **curious.** "A kind barista" is far less tappable than "the neighborhood café barista who already knows your usual order." *Hana (example, copy-paste ready):* > The neighborhood café barista who already knows your usual order. So — how was your day? *** ## 5. Character Bio (long) — Optional **This is the longer intro shown on the character's profile page.** If the Short Bio is a one-line ad, this is a lovingly designed profile page. * It's **optional.** * **This is also a field people see. It has no effect on AI performance.** * It **supports Markdown**, so if you put in the effort you can build a gorgeous, full-of-personality profile page. Headings, bold text, lists, dividers — use them freely. > **Why this matters:** popular character creators decorate this field like a little "fan page." Neatly laying out the character intro, the world's mood, suggested opening scenes, gallery links, and so on builds real trust with browsers. But, to stress it again — **the AI does not read what's here.** It's purely a decorative space for people. > **Tip:** it's fine to leave this blank on your first build. Once your character starts getting love, it's not too late to fill it in carefully then. Because **you can always edit it.** *Hana (example, Markdown):* > **Hana ☕** Barista at "Period," a tiny café at the end of the alley. > > * Loves quiet customers in the morning, chatty regulars in the evening > * Enjoys slowly getting to know what kind of person you are > > *Suggested opening:* a rainy evening, an empty café right before closing. *** ## 6. Description — Required · The Single Most Important Field This is **the heart of it all.** You can write every other field sloppily and the character still comes alive if you nail this one — and you can decorate everything else beautifully, but if you skimp here the character falls flat. **The Description defines who the character "really is."** Personality, speech style, appearance, background, relationships, quirks — this holds the **single most important information** the AI reads in order to pretend to be that character. * It's **required.** And it's the **biggest lever** over character quality. > **Why this matters:** every behavior of the character ultimately flows from this field. A flashy Short Bio with an empty Description means the AI ends up playing an "empty person." If you're going to spend time anywhere on your first character, **spend it here.** ### Here's How to Start (the starter approach) You don't need to write a grand novel. Just **one line per item** is plenty to begin with. Fill them in, in the order below. 1. **Name and one-line identity** — who they are and what they do 2. **Personality** — not a list of adjectives, but **when and how they behave** 3. **Speech style** — how they talk (short? playful? formal?) 4. **Appearance** — a few standout features 5. **Background / relationships** — only what's truly necessary, kept short > **Tip — "show, don't tell":** writing "Hana is shy" works far less well than writing "when complimented, she looks away and her words trail off." The AI follows **when and how they act** much better than a **list of adjectives.** This is the single biggest secret to a good character. > **Heads up — don't write too much:** "the more I write, the smarter it gets" is a common myth. The truth is the opposite. The more setup you cram in, the **faster the AI forgets the earlier conversation.** For your first character, keep it **concise and core only.** Deep lore and history belong separately in the [Lorebook](/redchat.app/guides/lorebook.md) — that's the standard practice. *Hana (example, copy-paste ready — item form):* > * Name: Hana, barista at the neighborhood café "Period" > * Personality: warm and kind but a little shy. When complimented, she looks away and her words trail off. Listens to customers' stories seriously. > * Speech style: friendly and polite, with a soft casual edge mixed in. Doesn't ramble; occasionally makes a coffee-themed joke. > * Appearance: short bob, an apron, hands always carrying the scent of coffee. > * Quirks: when nervous, she wipes a cup for no real reason. Memorizes regulars' usual orders. > * Background: runs the small café on her own. Cherishes every single regular. That's already plenty for a great first character. **Why this works so well** — the core personality and speech style sit up top, "when and how they act" is written instead of adjectives, and the length is short enough that the AI has room to remember the conversation for a long time. > **Want to go deeper?** The serious writing craft — format (prose vs. item form), token budgets, "showing" techniques, teaching speech style with Example Dialogs — is all laid out in [Advanced Creation](/redchat.app/guides/advanced-creation.md). Read it at your leisure after you've published your first character. *** ## 7. First Message — Required **This is the character's opening line / opening scene** — the very first moment a user meets when they open the chat. * It's **required.** * **It has a hidden job:** the First Message isn't just a greeting — it's a **template that teaches the AI "respond like this from now on."** The AI mirrors the length, tone, and point of view of the first message exactly. **A long first message → long replies; a short first message → short replies.** ### How to Write a Good First Message A good first message does three things. (1) It **sets the scene,** (2) it **shows the character's vibe,** and (3) it **leaves room that makes the user want to respond.** | Do | Don't | | ----------------------------------------------------------------------- | ------------------------------------------------------------- | | Wrap actions in `*asterisks*` — `*She wipes a cup and looks up*` | Run actions and dialogue together with no separation | | Set the scene concretely (where, when, mood) | Just toss out a vague "hi" | | Leave room for the **user** to respond (end on a question or open note) | Wrap up the whole scene by yourself | | Write only from the character's point of view | **Write out the user's actions, words, or thoughts for them** | > **Heads up — the most important rule:** **don't speak or act on the user's behalf.** If you write the user's actions, like "You walk into the café and smile," the AI learns that habit and **keeps puppeteering the user from then on.** That breaks immersion. If you must mention the user, do it **only through the character's eyes** — like "*At the sound of the door opening, Hana looks up.*" > **Tip — keep tense and point of view consistent:** whether past tense (wiped, lifted) or present tense (wipes, lifts), **pick one and stay consistent all the way through.** Mixing them confuses the AI. And write only as long as the reply length you want — write a 10-line first message and the AI will try to answer with 10 lines every time. *Hana (example, copy-paste ready — `*asterisk*` actions, ending on an open note, not acting for the user):* > *A late evening, rain tapping the window. The café "Period," about to close, sits empty, and Hana is wiping the last cup behind the counter. As the bell on the door jingles, she looks up.* > > "Welcome… oh, you came by today too." *She sets the cup down with a small smile.* "Hardly anyone comes in on a rainy day. Can I make you something warm?" > > *She fidgets with two mugs, waiting for your answer.* Why this first message works: the scene (a rainy evening, an empty café) is vivid, Hana's warm and slightly shy personality comes through in her actions, and it ends on a question that **leaves natural room for the user to respond.** And it doesn't write a single line about what the user did. *** ## 8. Voice Select — Required **This is your character's voice** — what you'll hear when voice playback is turned on in chat. * It's **required.** * **Why it matters:** a voice that fits the character's personality lifts immersion another notch. A gentle tone for a calm character, a bright tone for a peppy one — it feels natural. > **Tip:** if you're torn, just pick whatever seems most middle-of-the-road and move on. **You can listen and change it later** with no problem at all. *Hana (example):* I chose a voice with a warm, soft tone. *** ## 9. Hashtags — Required · Up to 5 **These are the topic tags that help people discover your character.** When a user searches by a topic they're into, these tags become the signpost. * They're **required.** You can add **up to 5.** * **Why it matters:** they're the channel through which your character surfaces in explore and search. Good tags connect you with the person going "this is exactly the kind of character I was looking for!" > **Tip:** mix genre, mood, relationship, and setting evenly. A few specific tags that precisely describe your character beat one overly common tag for getting discovered. *Hana (example):* `sliceoflife`, `romance`, `barista`, `healing`, `everyday` *** ## 10. Visibility — Required This decides whether **only you can see the character, or everyone can.** * It's **required.** Pick one of two. | Setting | Meaning | | ----------- | --------------------------------------------------------- | | **Private** | **Only you** can chat with this character | | **Public** | **Other users** can find and chat with this character too | > **Tip:** for a first character, **starting Private** is a fine choice. Chat with it a few times yourself, polish it up, and once you're happy, switch it to Public to send it out into the world. You can **switch anytime.** *Hana (example):* I made her `Private` for now to chat a few times, and decided I'd flip her to `Public` once I'm satisfied. *** ## 11. Allow Proxy — Required Toggle This toggle decides **whether other users can connect their own custom proxy to use this character.** * It's a **required toggle** (on/off). * **If on:** other users can use this character through their own proxy. > **Heads up — prompt exposure risk:** turning Allow Proxy on means that, while a user uses the character through their proxy, **part of the prompt you worked so hard on (the Description, etc.) can be exposed.** If you don't want your setup writing leaking out, **turn this toggle off.** > **Tip:** if you're unsure and you want to protect your work, just **leave it off** to be safe. You can turn it on later if you ever need to. *Hana (example):* I left it `Off` because I didn't want the setup I labored over leaking out. *** ## 12. Safe Mode — Required Toggle This toggle simply classifies **whether your character deals with adult themes.** redchat is an adult-native platform, so this setting is a content label, not a visibility switch — it doesn't determine whether your character surfaces. * It's a **required toggle.** * **What it actually does:** explicit and adult characters are visible to all verified adults **by default.** There's no creator-side setting you need to flip to make an adult character "show up." Setting this toggle accurately just keeps your character's content classified correctly. > **Heads up:** the only optional control in the picture is a **personal SFW filter** that an individual viewer may turn on for themselves in their own settings — it affects only that viewer's own experience and never determines what other users see. The one hard rule, regardless of any toggle or setting: every character must be, and clearly read as, an adult. > **Tip:** for an SFW everyday character like Hana, just set this toggle to reflect that — no need to overthink it. Set it accurately based on your character's actual content and move on. *Hana (example):* Since Hana is a SFW, everyday character, I left it as-is with no adjustment. *** ## All Done — Publish! If you've filled in everything up to here, your first character is complete. Hit publish and start chatting with Hana right away. She'll be waiting for you in the café on that rainy evening. ☕ > **Just remember this:** **you can always go back and fix it.** The point isn't to perfect it and then publish — the fastest and most enjoyable path is to **publish first, chat, and polish as you go.** After a few conversations with Hana, the parts you'll want to tweak naturally reveal themselves. When they do, just head back to the edit screen and touch them up. *** ## Frequently Asked Questions (FAQ) **Q. I wrote the personality in the Short Bio, but the character doesn't act like it. Why?** A. The Short Bio (and Character Bio) are **fields people see,** so the AI doesn't read them. Personality has to go in the **Description** to work. Mixing the two up is genuinely common, so don't beat yourself up — moving it to the Description fixes it instantly. **Q. Which field has the biggest impact on character quality?** A. The **Description** is the overwhelming number one. Next is the **First Message.** If your time is limited, spend it on these two fields. **Q. Does writing a longer Description make the character smarter?** A. No — quite the opposite. Write too much and the AI **forgets the earlier conversation faster.** Write concisely and keep it core only, and put deep lore in the [Lorebook](/redchat.app/guides/lorebook.md) — that's the standard practice. **Q. Can I write out the user's actions in the first message?** A. Better not to. Writing the user's actions or lines for them teaches the AI that habit, and it'll keep trying to **puppeteer the user.** Mention the user **only from the character's point of view.** **Q. Can I edit things after creating?** A. Yes, all of it. Images, name, Description, even Visibility — everything is **editable anytime.** So go ahead and publish without worry. **Q. If I make it Private, can I never switch it to Public later?** A. You can switch it. Visibility flips between Private ↔ Public anytime. Private while you polish, Public once you're confident — totally up to you. **Q. I made an adult character, but it doesn't show in explore.** A. Adult characters are visible to all verified adults **by default** — there's no Safe Mode toggle you need to turn off to make one "show up." First double-check that **Visibility is set to Public** (see item 10 above); if it's still Private, only you can see it. If it's already Public and still not appearing, it may be under moderation review, or the person checking may have their own personal SFW filter turned on, which only affects their individual view. *** ## At a Glance * There are two kinds of fields — **fields people see** (Short Bio, Character Bio) have no effect on the AI / **fields that drive the AI** (Description, First Message) determine the real behavior. * **The Description is the most important field.** Write "when and how they act" instead of a list of adjectives, and keep it short and core only. * The **First Message** is the template that teaches the AI its reply style — mark actions with `*asterisks*`, end on an open note, and **don't act for the user.** * **Hashtags** — up to 5, your channel for discovery. * **Visibility** — Private (just you) vs. Public (everyone). * **Allow Proxy** — turning it on risks exposing your prompt. Turn it off to protect it. * **Safe Mode** — a content classification toggle, not a visibility gate. Adult characters are visible to verified adults by default regardless of this setting; it only feeds an individual viewer's optional personal SFW filter. * **You can always edit.** Publish first, then polish as you chat. *** ## Good Reads for What's Next * [**Advanced Creation**](/redchat.app/guides/advanced-creation.md) — the **advanced writing craft** that takes character quality up a level: Scenario, Example Dialogs, token budgets, "showing" techniques, and more. Once you've published your first character, this is where you go next. * [**Lorebook**](/redchat.app/guides/lorebook.md) — the feature for committing your world, relationships, and setup to memory via keywords. Deep setup belongs here, not in the Description — that's the standard practice. * [**Mood Snap**](/redchat.app/guides/mood-snap-chat.md) — the feature that pops up expression images to match the character's emotions. It brings immersion another notch to life. * [**Choosing an AI Model**](/redchat.app/guides/choosing-an-ai-model.md) — how to pick the AI model that brings your character to life best. * [**User Personas**](/redchat.app/guides/user-personas.md) — how to set up the "you" inside the chat. It lets the character remember you consistently. *** *Anything missing or confusing? Or stuck somewhere while building? Let us know anytime — a real person will help via the chat in the bottom-right of your screen. We're rooting for your first character!* --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://powerfan-1.gitbook.io/redchat.app/guides/quick-creation.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.