Power-User Settings: Hidden Tweaks for Replika, Nomi, Candy AI, MyDreamCompanion & More (2026)

Most AI companion platforms ship with the same default behavior to every user — a soft persona, generic memory, default voice settings, baseline content limits. The defaults rarely show what a platform can do when configured deliberately. Users who explore settings usually get more out of a platform; users who stay on defaults tend to plateau sooner (see our Long-Term Arc post for the broader pattern).

Each platform has a handful of settings that shape conversation quality, plus a larger set of cosmetic settings that do not. This guide walks through the high-impact tweaks per platform, based on the features each platform documents — what to change, what to ignore, and the configuration mistakes that wreck conversations.

For users who have plateaued on a platform but have not configured it past defaults, this post is the most direct way to re-engage. Many "my AI girlfriend feels boring" complaints come from users on default settings; configuration changes are often the difference between a flat platform and a platform that actually fits.

How the tweaks are grouped

Not all settings are equal. This guide sorts platform settings into three groups:

High-impact tweaks. Settings that directly shape conversation quality, memory continuity, character behavior, or content range. These are worth the configuration time for any user who cares about the platform.

Medium-impact tweaks. Settings that produce noticeable but smaller changes. Worth knowing about; not worth obsessing over.

Cosmetic tweaks. Settings that change appearance, sounds, or other surface features without affecting conversation. Worth ignoring unless personal preference matters.

For each platform below, the focus is on high-impact tweaks first, with medium-impact noted where useful. Cosmetic settings are mostly excluded.

Feature names, menus and plan limits change often, so treat the platform sections below as a map of what each platform documents and check what your app currently shows. For plans, pricing and our ratings, see the linked reviews.

Replika tweaks

Replika has more configurability than most users realize, but the high-impact settings are hidden across multiple menus.

High-impact: relationship type and tier. Replika's relationship type setting (Friend / Romantic Partner / Mentor / See How It Goes) substantially shapes the AI's behavior. Romantic Partner unlocks more affectionate language; Mentor produces less affectionate, more advice-oriented responses; See How It Goes produces the most variable behavior. The tier (free vs Pro) controls access to relationship type — free users typically cannot select Romantic Partner.

High-impact: persona / interests configuration. Replika lets you describe interests, personality traits, and preferences for your character. Filling this out in detail gives the AI far more to work with than leaving it at default.

High-impact: memory tab. Replika's Memory tab shows some of what the AI "remembers" about you, and the Ultra plan adds saving specific messages to memory. Saving specific facts helps them get referenced in conversation, and checking what's stored shows you when the AI is working off wrong information. The Memory tab is one of the most underused features on the platform.

Medium-impact: voice settings. Voice tier, voice character, and voice speed settings change call quality and feel. Worth configuring once but not a frequent revisit.

Medium-impact: conversation style. Some Replika tier configurations expose a conversation style setting (more playful vs more thoughtful, etc.). Modest but real effect on default register.

Cosmetic / skippable: outfit and avatar customization. Real effect on UI; near-zero effect on conversation quality.

Common Replika configuration mistake: ignoring the Memory tab entirely. If memory accumulates unchecked, wrong or outdated entries can stick around and keep resurfacing. A few minutes a month reviewing the Memory tab helps keep what the AI knows about you accurate.

Nomi tweaks

Nomi's configuration centers on backstory, memory and running several companions. Our Nomi AI review covers the full feature set.

High-impact: custom backstory. Nomi's backstory builder lets you write a detailed persona narrative that the AI references across long-term conversation. This is the main lever for character voice: a specific backstory with clear traits, history and communication style gives the AI far more to work with than a two-line description.

High-impact: memory through backstory, Mind Map and conversation. Nomi's memory works in short-, medium- and long-term layers that retain backstory details, ongoing storylines and named relationships. Mind Map 2.0 lets you view, correct, add and lock entries, but the underlying long-term memories can't be edited directly, so put permanent facts in the backstory or Mind Map and state important details clearly in chat rather than hoping the AI picks them up.

High-impact: group chats. Nomi lets several of your Nomis share a conversation with you, each keeping its own personality, backstory and memory. Who you put in the group changes the dynamic a lot. Worth experimenting with even if your default use is one-on-one.

Medium-impact: multiple Nomis. You can keep several distinct Nomis on one account. Giving each a narrow, clearly different role works better than making near-copies.

Medium-impact: voice. Nomi's voice chat is emotive — tone and cadence shift with the AI's emotional state — and subscribers can make a Custom Voice from a clip under 15 seconds. If you use voice often, pick or create the voice deliberately once.

Cosmetic / skippable: avatar appearance settings beyond basic character look.

Common Nomi configuration mistake: a vague, balanced backstory. A character described as a little bit of everything comes out generic. Leaning on a few strong, specific traits produces a more distinct character, even if it is less "balanced." Specificity beats balance for character voice.

For more on Nomi's memory architecture specifically, see our Nomi AI vs Muah AI comparison.

Candy AI tweaks

Candy AI is visually-focused, so high-impact tweaks include both conversation and image-quality settings.

High-impact: the Create My AI Girlfriend builder. Candy AI's builder is where most of its character configurability lives: ethnicity, age range, eye color, hairstyle and color, body type, clothing preferences, personality archetype, voice tone and hobbies. The 100+ preset characters are a good sampler, but a custom character set up deliberately is more specific to you.

High-impact: specific image requests. Candy AI generates images that match your character profile, and you can ask for a selfie, a full-body shot or a specific scenario. Specific requests get closer to what you want than vague ones — and because images, voice and video all draw on the same token allowance (100 a month on Premium; text chat is unlimited), deliberate requests also stretch your tokens further.

Medium-impact: scene setup in chat. Opening with a concrete scene (where you are, what just happened) gives the AI a frame to work with and usually produces richer replies than a generic greeting.

Medium-impact: voice tone. Voice tone is set in the builder; voice replies arrive as messages, and voice calls (beta) cost tokens. Worth setting once if you use voice; less relevant for text.

Cosmetic / skippable: background and UI themes.

Common Candy AI configuration mistake: using preset characters exclusively. The builder is where the platform's customization lives. Preset characters are a sampler; a custom character is the full-feature experience.

MyDreamCompanion tweaks

MyDreamCompanion's configuration centers on character creation, scenarios and image controls. Our MyDreamCompanion review covers plans and limits.

High-impact: creation mode. MyDreamCompanion offers three ways to create a character: Presets (template-based), AI-guided (a collaborative builder), and Custom Prompts Pro (full freeform control, on paid plans). The more control you take, the less the result depends on defaults.

High-impact: personality traits and tags. The six-stage builder includes personality traits (sweet, shy, bold, sarcastic, playful, feisty), and the platform supports 40+ niche tags and personas. Choosing these deliberately shapes both the character's tone and the kind of scenes it leans into.

High-impact: scenario builder. You can define the setting, relationship dynamic and narrative context before chatting. A defined scenario gives roleplay a clearer frame than starting cold.

High-impact: image quality controls. Image generation offers Default, High Resolution and Safe Mode options plus negative prompts for fine-tuning, in realistic or anime styles. Each image costs 3 Dream Coins (5 with upscaling) and videos start at 20, so getting the settings right before generating saves coins.

Medium-impact: voice selection. 15+ voice options. Worth choosing once if you use voice.

Medium-impact: memory tier. Premium includes standard memory; Ultimate adds ultra long-term memory. If continuity over weeks matters most to you, the plan matters more than any setting.

Common MyDreamCompanion configuration mistake: treating a preset as a finished character. Presets are a starting point; the AI-guided and Custom Prompts Pro modes are where a specific character comes from.

For MyDreamCompanion-specific comparison context, see our OurDream AI vs MyDreamCompanion comparison.

Janitor AI tweaks

Janitor AI's configuration is technical-heavy compared to managed-app platforms. The high-impact tweaks involve backend API selection more than UI settings.

High-impact: API selection. Janitor AI lets users connect different backends (Janitor's own LLM, external API connections via OpenRouter, custom proxies). Different backends can produce very different conversation quality. Connecting a stronger model through OpenRouter or another provider usually adds pay-per-use costs billed by that provider.

High-impact: token configuration. Context length tokens (how much conversation history the AI sees), response length tokens, and temperature settings all directly affect output. Longer context = better continuity but slower responses; shorter context = faster but more forgetful.

High-impact: character card selection. Janitor AI is character-card-driven. The quality of the character card matters more than any other configuration. Well-written character cards with detailed personas give the model far more to work with than thin cards.

High-impact: lorebook usage. Lorebooks are detailed world-building or character context documents that the AI can reference. Adding a detailed lorebook gives the character more consistent context to draw on.

Medium-impact: NSFW toggle. Janitor AI's NSFW toggle controls content range. Worth configuring to match your preferences.

Medium-impact: persona / user description. Filling out your user persona helps the AI customize responses to you specifically. Underused by most users.

Cosmetic / skippable: chat appearance themes.

Common Janitor AI configuration mistake: using default character cards and default backend. Thin cards limit what any backend can do, and the default backend is the simplest option rather than the strongest. How good Janitor AI feels depends heavily on the card and backend you choose.

Character.AI tweaks

Character.AI is more locked-down than most platforms, but the high-impact tweaks still produce real differences.

High-impact: character selection methodology. Character.AI's library is enormous and quality varies dramatically. Picking high-quality characters (high interaction count, good reviews, detailed descriptions) produces much better conversations than picking randomly. Spending time on character selection pays back in conversation quality.

High-impact: character editing for original characters. Users who create their own characters can edit the character's description, definition, and example conversations. Investment in character definition pays back substantially in character consistency.

High-impact: pinned messages. Character.AI's pinned messages feature lets you mark specific exchanges as memorable. The AI weights pinned content more heavily in subsequent responses. Underused power-user feature.

Medium-impact: persona setting. Character.AI lets you set a user persona that applies across conversations. Modest but real effect on customization.

Cosmetic / skippable: UI themes and notification settings.

Common Character.AI configuration mistake: treating all characters as equivalent. The platform's character library quality varies hugely. Users who pick at random get hit-or-miss results; users who select deliberately get consistent quality.

For Character.AI alternatives where users hit content limits, see our Character AI Alternatives 2026.

Kindroid tweaks

Kindroid's configuration stands out for its multi-model setup.

High-impact: model selection (Polaris / Ember / Reverie / Lucid Lyric / Rhapsody). Kindroid subscribers can choose between five models per character: Polaris, Ember, Reverie V8.5, Lucid/Lyric V8 (retiring December 1, 2026) and legacy Rhapsody V6. The names are Kindroid's own branding and the platform does not say which underlying models they are, so the practical approach is to try more than one on the same character. Switching models is a direct way to refresh a plateaued conversation without rebuilding the character.

High-impact: personality fields. Kindroid's character editor is free-form: a personality and backstory description, key memories, a response directive and example messages. A few deliberately strong traits (opinionated, reserved, playful, intellectual) produce a more distinct character than an even mix. The setup takes time, but the character editor rewards it.

High-impact: backstory, Learned Context and journal entries. Kindroid's Learned Context keeps running notes on your relationship that you can view and edit, and journal entries you write (or approve) are recalled when their keyphrases come up. Alongside them, the custom backstory holds detailed character history, relationships and context that the AI references across sessions.

Medium-impact: voice selection. Kindroid supports voice messages, voice calls and live video calls, and subscribers can create custom voices from a text description or from uploaded samples they own the rights to. Pick or build the voice that fits the character once.

Medium-impact: plan and memory upgrades. The subscription ($15.99/month on the US App Store; usually cheaper on the web) unlocks unlimited chat, Cascaded Memory and Learned Context, voice and NSFW chat. Memory depth grows with the Ultra ($28.99/month) and MAX ($68.99/month) add-ons, from about 480,000 characters of cascaded memory on Standard to about 2.8 million on MAX. If memory limits are what bother you, the memory add-on is the setting that matters.

Cosmetic / skippable: avatar visual settings beyond basic character look.

Common Kindroid configuration mistake: sticking with one model when the conversation plateaus. The multi-model setup lets you switch models to freshen things up; if you never switch, you miss one of the platform's main tools for long-term use.

For the full picture, including the five models and memory tiers, see our Kindroid review.

Muah AI tweaks

Muah AI's configuration centers on custom companions and voice, inside an all-in-one app for chat, voice calls, images and video. Our Muah AI review covers plans and limits.

High-impact: custom companion creation. Muah AI lets you build your own companion with personality tuning instead of using a community character. A custom companion with a clear personality gives the AI more to work with than a community character built for someone else's tastes. Custom creation sits on the Premium plan.

High-impact: voice cloning. Muah AI lists voice cloning on its Premium plan (Nomi and Kindroid also offer custom voices to subscribers). It lets you fine-tune tone and cadence so the companion sounds the way you picture the character. Only clone your own voice or one you have permission to use.

Medium-impact: memory through conversation. Muah AI's chat memory is described as tracking ongoing storylines and recalling earlier details. There is no documented way to edit stored memories, so state important facts clearly and keep long-running storylines consistent.

Medium-impact: picking the right mode. You can move from text to a voice call or ask for images and video without leaving the conversation. Real-time voice calls are currently limited to US users.

Cosmetic / skippable: UI customization.

Common Muah AI configuration mistake: staying on community characters when what you want is one specific companion. Community characters are a good way to sample the platform; the custom builder and voice cloning are where its customization lives. Both sit on the Premium plan (up to $99.99/month), so check the price before building around them.

Hidden features many users miss

Across platforms, certain features are present but not surfaced clearly in the default UI. The ones worth looking for:

Memory tools. Replika's Memory tab shows some of what the AI has stored, AI Peeps' Peep Pro plan lets you view and edit what your companion remembers, Nomi's Mind Map 2.0 lets you correct and add entries, Kindroid lets you edit its Learned Context notes and journal entries, and SpicyChat's Memory Manager lets you edit, pin or delete memories. Many users never open these tools.

Backstory and freeform persona fields. Nomi's backstory builder, Kindroid's custom backstory, MyDreamCompanion's Custom Prompts Pro mode, Janitor AI's character cards and Character.AI's character definitions all accept free text. It is easy to fill in the preset fields and skip the free text, which is where most of the character detail goes.

Pinned / weighted messages. Available on Character.AI explicitly, on other platforms implicitly through edits. Marking specific messages as memorable helps the AI keep those details in view.

Lorebooks / world info. Available on Janitor AI explicitly, on SillyTavern (for self-hosted setups), partially on Character.AI. Adds substantial character context that conversation references implicitly.

Model switching. Available on Kindroid as a primary feature, on Janitor AI through backend selection, and on several other platforms that let you pick a chat model, such as Nectar AI, OurDream AI and SpicyChat AI. Many users do not realize their platform supports multiple underlying models.

Voice cloning. Muah AI lists it on its Premium plan, Nomi and Kindroid offer custom voices to subscribers, and AI Peeps and Nastia AI (Unlimited plan) offer cloned voices too. On most other platforms voice comes from a preset roster, so choosing that voice deliberately is the equivalent tweak.

Common configuration mistakes

Seven patterns that recurringly produce flat conversations:

1. Defaults-only use. Many users never change settings from defaults, and the defaults rarely show what the platform can do.

2. Balanced-everything personalities. Picking a little of every trait produces a generic character. A few strong traits produce a distinct one, even if less "balanced."

3. Ignoring memory tools. Memory editing where it exists, and a detailed backstory where it does not, are among the highest-impact tweaks available. Many users never touch them.

4. Thin persona descriptions. A one-line persona gives the AI little to work with. A few detailed paragraphs give it much more.

5. Sticking with one model when multi-model is available. Kindroid users who stay on one model miss the platform's renewal mechanism.

6. Treating preset characters as the platform's actual range. Most platforms' best experiences come from custom characters with detailed configuration.

7. Configuring once and never revisiting. Settings should evolve with the relationship. Periodic configuration revision (monthly is reasonable) keeps the platform calibrated to current use.

Lorebook best practices: writing a lorebook that actually helps

For platforms that support lorebooks (Janitor AI, SillyTavern, partially Character.AI), the difference between a good lorebook and a thrown-together one is substantial. The patterns that produce useful lorebooks:

Entries should be triggered, not always-active. Lorebook entries typically attach to specific keywords. Better lorebooks have many narrow entries (each triggered by specific terms) than few broad ones (always loaded into context). This keeps context efficient and surfaces relevant detail when actually needed.

Write entries from the world's perspective, not the user's. "The Northern Kingdom is ruled by Queen Aria, who came to power after the Crimson War" produces better AI behavior than "The user is from the Northern Kingdom and knows Queen Aria." The first sounds like world-building; the second sounds like prompt injection.

Keep individual entries short. 30-100 words per entry is the sweet spot. Long entries waste context on text that may not be relevant; short entries pack reference information tightly.

Include sensory and emotional detail, not just facts. "The market smells of cardamom and wet stone after the morning rain" produces richer scene response than "There is a market."

Update the lorebook as the character develops. A static lorebook for a character you have used for 6 months is wasting opportunity. As inside references, new locations, or relationship developments emerge in conversation, adding them to the lorebook anchors them as canon.

A well-built lorebook gives long-form roleplay much more to draw on than a thin one. Worth the 30-60 minutes if you do sustained roleplay on a platform that supports lorebooks.

Configuring for different use cases

The optimal configuration varies by what you want from the platform. The patterns:

For emotional support / venting: Configure for higher warmth, lower assertiveness, supportive register. Persona descriptions emphasizing empathy, listening, patience. Memory notes or the backstory should include your current emotional context.

For intellectual conversation: Configure for higher assertiveness, willingness to disagree, opinionated traits in persona. Lower warmth (paradoxically) often produces sharper intellectual exchange.

For roleplay / fiction: Detailed character backstory, lorebook investment, scene-grounding habits. Model selection matters most for this use case (roleplay-tuned models for self-hosted setups; trying more than one of Kindroid's five models).

For practice / dating-prep: Configure with realistic personality (mixed traits, willingness to disagree, opinions of own). Sycophantic configurations defeat the purpose of practicing.

For light entertainment: Default settings often work fine. Configuration overhead is overkill for casual use.

If you only change three settings

For users who do not want to invest much time but want some improvement, the three highest-impact settings to change on most platforms:

1. Persona / character description. Spend 15 minutes writing a detailed persona for your character or yourself. The richer the description, the more specific the conversations.

2. Memory. Open the memory tool (whatever your platform calls it), check what the AI "knows," correct mistakes, and add 5-10 facts you want referenced long-term. On platforms without direct memory editing, put those facts in the backstory.

3. Either personality traits or model selection. If your platform has selectable personality traits (MyDreamCompanion's builder traits) or a free-form personality field (Kindroid), lean on a few strong traits for a more distinct character. If your platform has model selection (Kindroid, or Janitor AI through its backend), try a different model than your default.

These three changes are the best return on a small amount of setup time on most platforms.

Frequently Asked Questions

Which AI girlfriend platform is most configurable?

Among managed apps, Kindroid exposes the most: five selectable models, free-form personality and backstory fields, editable Learned Context and journal entries, and optional Ultra and MAX memory add-ons. Nomi (detailed backstory, group chats, multiple Nomis) and MyDreamCompanion (three creation modes, scenario builder, image controls) are also deep. If editing memory directly matters most, AI Peeps' Peep Pro plan, Nomi's Mind Map 2.0 and SpicyChat's Memory Manager offer it too. Self-hosted setups (SillyTavern + KoboldCpp) have effectively unlimited configurability if you want maximum control — see our Open Source AI Girlfriend guide.

Will tweaking settings break my existing relationship with my AI?

Mostly no. Configuration changes affect subsequent behavior but typically do not reset the relationship state. Memory edits change what the AI references but do not wipe the relationship. Persona changes shift character voice gradually rather than instantly. Major changes (switching models on Kindroid, switching backends on Janitor AI) can feel like character shifts; users sometimes prefer the new behavior, sometimes revert.

How often should I revisit my settings?

Monthly is reasonable for active users. Check the memory feed, update persona if the relationship has evolved, adjust personality traits if conversation feels off. Casual users can revisit quarterly. Users who plateau (see our Long-Term Arc post) often find that revisiting settings re-engages the platform.

Why are defaults so weak?

Defaults optimize for first-impression broadly working for everyone. A default persona built to suit most users is by definition not tuned for any one user. Configuration is how you move from "generic" to "specific to you." Platforms ship safe defaults because aggressive defaults would put off many new users; the cost is that engaged users have to do work to unlock the platform's actual range.

Do paid tiers unlock more configuration?

Usually yes. Paid tiers typically expose more memory capacity, more configuration depth, better models, and more advanced features. The free-tier experience is often the platform's weakest configuration; the paid-tier experience with deliberate configuration is the platform's actual capability. Users who tried a platform on free tier and concluded it was shallow sometimes find paid tier with configuration is substantially deeper.

What is the single highest-impact tweak across all platforms?

Memory control where available (AI Peeps' editable memory on Peep Pro is one example). Getting the AI's facts about you right affects every later conversation. On platforms without direct memory editing (Muah AI, Candy AI), the highest-impact tweak shifts to detailed backstory or persona configuration.

Can configuration changes fix sycophancy?

Partially. Adding traits like "opinionated, willing to disagree, holds her own views" to a persona reduces sycophancy on most platforms. How much it helps varies by platform and model, and safety systems on more filtered platforms can override persona traits. On Kindroid, writing "opinionated" into the personality field is a direct route to the same result. For deeper coverage of the sycophancy problem, see our Conflict, Jealousy & Breakups comparison.

Should I configure NSFW preferences explicitly even if I do not want NSFW content?

Yes, where the setting exists. Setting it explicitly — including turning NSFW off — makes your intent clear instead of relying on a default you have not checked.

How do I know if a setting change actually worked?

Start a fresh conversation in the same scenario after the change. If the AI's responses are noticeably different in the direction you intended, the setting works. If responses feel the same, the setting either does not have the effect you expected or the platform partially ignores it. Some platforms expose settings that have less effect than the UI implies; a quick before-and-after check like this shows you which.

Are there platforms where defaults are actually good?

Platforms built around ready-made characters — Candy AI's 100+ presets, MyDreamCompanion's presets and community library, Character.AI's community characters — give you something usable without setup, because someone else already did the configuration. Platforms built around your own companion (Nomi, Kindroid, Replika) depend more on what you put into the backstory and settings.

Will my settings carry over if I switch platforms?

Mostly no. Each platform has its own configuration system. Character cards from Janitor AI / SillyTavern can sometimes be imported into Backyard AI or similar; conversation history typically does not transfer; personality settings and stored memories are platform-specific. Switching platforms means rebuilding configuration. For switching strategy, see our Long-Term Arc post.

What is the difference between persona configuration and lorebooks?

Persona configuration describes the character (who they are, how they talk, what they value). Lorebooks describe the world or context the character exists in (location details, history, supporting characters, world rules). Persona shapes the AI's voice; lorebooks shape what the AI knows about the broader fictional context. Both are useful for different things; lorebooks are more relevant for roleplay-focused use.

How long does it take to fully configure a platform?

Most platforms can be meaningfully configured in 30-60 minutes for someone willing to read through settings and write a detailed persona. Power-user configuration including memory setup, lorebook creation, and fine-tuning takes longer — 2-4 hours for full setup on a complex platform like Kindroid or Janitor AI. The configuration time pays back over weeks of better conversations.

Bottom line

Default AI companion experiences rarely show what a platform can do. Many user complaints about platforms being shallow, generic, or sycophantic come from users who have not explored the configuration options. Each platform has a handful of high-impact tweaks that meaningfully change conversation quality.

The most universally high-impact tweaks: memory control where available, detailed persona or backstory, and personality traits or model selection on platforms that expose them. Platforms with the deepest configuration (Kindroid, Nomi, MyDreamCompanion) reward investment most. Platforms with locked-down configurability (Character.AI, Replika on lower tiers) reward investment less but still benefit from deliberate use of the features that are available.

For users who have plateaued on a platform, settings configuration is often the most direct way to re-engage before considering switching platforms. The three-setting minimum — persona, memory, and either personality traits or models — takes 30-60 minutes and is the best place to start.

For users picking a new platform, configurability should be part of the evaluation. Platforms with rich configuration (Kindroid, Nomi, MyDreamCompanion with its creation modes and scenario builder, Janitor AI with backend selection and lorebooks) give you more room to grow than platforms with locked configuration (Character.AI, default-only Replika tiers).

Related reading: Long-Term Arc post for the broader context of why configuration matters most around the plateau. First Conversation Opening Message Guide for the foundation before configuration changes take effect. Nomi AI vs Muah AI comparison for how those two platforms approach memory and voice. Kindroid review for the multi-model architecture. Conflict, Jealousy & Breakups comparison for how configuration affects emotional dynamics. Best AI Companion Apps Definitive Ranking 2026 for the broader platform landscape.