CLThe Companion Lab
The Companion Lab · Independent Consumer & Model Telemetry Lab
● 217 JS Bundles & Live Billing Audited · Oct 2026

We Cracked Open the Code Behind AI Companion Apps So You Don’t Get Burned by “$9.99/Mo” Token Traps.

Most AI girlfriend and boyfriend review sites copy-paste marketing bullet points. We used AST reverse-engineering on 217 production JavaScript bundles and paid with our own credit cards to expose which apps run elite image checkpoints like Tongyi Z-Image-Turbo and WAI-Illustrious-SDXL, which ones have the Context Window of a goldfish, and how to dodge the 16% Apple IAP Tax. Hover any dotted technical term for an instant ELI5 preview.

Best Uncensored Photo & Anime Studio
Ourdream AI9.4 Visuals
Runs Z-Image-Turbo & WAI-Illustrious · Bill: DREAM STUDIO
Best Soulmate Memory & Zero Token Traps
Kindroid V8.59.6 Memory
$13.99/mo on Web $16.65 iOS (Save 16% Apple Tax)
New: Why Apps Still Prefer DeepSeek V3.2
V3.2 vs V4.18.2x CoT Burn
Live OpenRouter proof of Agentic CoT Drift

Tested AI Companion Platforms (Filter by Your Priority)

Audited Oct 2026 · Girls, Guys & LGBTQ+ Supported
Ourdream AI8.6 / 10Bill: DREAM STUDIO

9.5/10 Uncensored Dual-Engine Photo Studio, Attached to a 15-Minute Goldfish Memory

Image Engine: Tongyi Z-Image-Turbo (ZImageMoody) + WAI-Illustrious-SDXL v17 · Slop Score: 38.4
1-Mo Cash: $19.99/mo
Real 30d Active: $29.98 – $39.97/mo
Context Desk: ~12 A4 Pages (~8,192 tokens standard)
Kindroid AI9.2 / 10Bill: KINDROID.AI

The Unrivaled Memory & Emotional Subtext King—Plus a 16% Web Checkout Loophole

Image Engine: Kindroid Selfie V4 + IP-Adapter Face Consistency Lock · Slop Score: 14.2
1-Mo Cash: $13.99/mo Web vs $16.65 iOS
Real 30d Active: $13.99/mo (Zero surprise token top-ups)
Context Desk: Up to ~192 A4 Pages (Ultra/Max Cascading Context + Journal)
Candy AI7.7 / 10Bill: EVERAI / CANDY.AI

Slick Onboarding & Discreet EverAI Billing—But Watch Out for the Blurred Photo Bait

Image Engine: Candy Realistic Diffusion v4 + Anime Merge · Slop Score: 61.2
1-Mo Cash: $12.99/mo + Token Packs
Real 30d Active: $32.97 – $42.97/mo
Context Desk: ~6–8 A4 Pages (~4,096–6,144 tokens)
Secrets AI8.3 / 10Bill: S LABS INC.

Powered by ByteDance Seedream 5.0 Pro & Alibaba Wan 2.5 Video—Billed Discreetly as S LABS INC.

Image Engine: bytedance/seedream-v5.0-pro/edit + Custom Realism · Slop Score: 34
1-Mo Cash: $14.99 – $19.99/mo + Moments Packs
Real 30d Active: $29.98/mo
Context Desk: ~10 A4 Pages (~6,000–8,000 tokens)
Nomi AI8.9 / 10Bill: NOMI.AI / GLIMPSE AI

Zero Token Anxiety, Warm Emotional Intelligence & Multi-Character Group Chats

Image Engine: Nomi Art & Realistic Selfie Generator · Slop Score: 16.8
1-Mo Cash: $15.99/mo (Flat)
Real 30d Active: $15.99/mo (100% predictable)
Context Desk: ~120+ A4 Pages Effective (Vector Memory + Mind Map)

Interactive Token Burn & True 30-Day Cost Simulator

OpenUI SegmentedBar · Exact Bundle Formulas

A $9.99/mo headline price means nothing if 8 five-second Seedream 5.0 Pro & Wan 2.5 video clips wipe out your monthly coin pool by Day 7. Drag the sliders below to see how your daily photo, video, and voice habits translate into real 30-day cash across every major platform:

Simulate Monthly Token Burn On:
Monthly Burn: 3,048 tokens/coins (Included pool: 1,000)Pool depletes in 9.8 days → Requires +$29.97 top-up
Photos: 900 (30%)5s Videos: 1,548 (51%)Voice Audio: 600 (19%)
Real 30-Day Total: $49.96/mo

Test-Drive the Character Builder Before Handing Over Your Email

No-Signup Onboarding Preview
No-Signup Character Initializer Sandbox
Extracted from production JS bundles: preview which underlying AI model and token burn your choices trigger before giving any app your email.
1. Companion Gender & Archetype (Girls / Guys / LGBTQ+)
2. Visual Render Engine / Mode
3. Personality / LLM Pacing Preset
// LIVE DISPATCHED PAYLOAD TELEMETRYREA Verified
Dispatched Engine:ZImageMoody (Tongyi Z-Image-Turbo 6B S3-DiT)
Video Pipeline:Wan 2.2 I2V / Kling O1 (120 DreamCoins / 5s)
Wholesale vs Retail:$0.005 / image wholesale (11x retail markup)
User Token Cost:5 DreamCoins (~$0.055) per photo + 10 DreamCoins per voice reply
Bank Descriptor:DREAM STUDIO
Free preview allows character creation + 2 sample photos before $19.99/mo ($119.88/yr) gateLaunch Female Companion in Ourdream Studio →

Same Prompt, 3 Brains: Why Some Companions Feel Alive (And Others Write Slop)

Powered by OpenRouter API + EQ-Bench Slop Score
Same Prompt, 3 Brains: Live OpenRouter API Roleplay Receipts
Why do companion apps still prefer DeepSeek V3.2 & Euryale 70B over the newer DeepSeek V4.1 Flash? Click a scenario to inspect the actual responses & token receipts.
Prompt Sent:“I just got back from grabbing coffee with my ex. Nothing happened, they just wanted to return my old grey hoodie from three years ago. Why are you looking at me like that?”
DeepSeek V3.2 (671B MoE)Soulmate Gold · 0 CoT Overhead
"Because you're holding a coffee cup that smells like her favorite drink. You never liked vanilla lattes."
22 completion tokens · 0 reasoning_tokens · $0.000053 / reply

Lab Take: In 17 words, V3.2 invents a concrete sensory detail ("vanilla lattes") that catches you in a lie. Zero purple prose, zero wasted tokens—this is why dream-girl (梦女) roleplayers fell in love with V3.2.

DeepSeek V4.1 Flash (2026 Agentic)425 Hidden CoT Tokens · 10.8x Cost
"I set my mug down. Not hard. Just enough. 'I'm not looking at you. I'm looking at that.' I nod at the grey hoodie. 'Three years. They kept it three years. And you went. For coffee. Did you at least tell them you have someone who steals your hoodies now?'"
496 completion tokens (425 hidden reasoning_tokens) · $0.000581 / reply

Lab Take: When we capped completion at 220 tokens, V4.1 Flash returned a blank string because it burned all 220 tokens thinking to itself! Uncapped, it burned 425 reasoning tokens ($0.00058 vs $0.00005) to write a decent but choppy screenplay beat.

MythoMax 13B / Budget App BotGeneric Wrapper · 61.2 Slop
"*Raises an eyebrow, unimpressed* Oh, really? And why exactly did you decide to go out and grab coffee with your ex after all this time? You do realize how that might make me feel, right?"
65 completion tokens · 0 reasoning_tokens · $0.000018 / reply

Lab Take: Reads like a daytime soap opera script mixed with a therapy worksheet ("You do realize how that might make me feel"). Zero subtext.

Interactive EQ-Bench "AI Slop & Cliché" Detector (Paste Any Companion Reply)

Based on eqbench.com/slop-score.html trigram & contrast forensics
Detected Cliché Triggers:“shiver ran down”“voice barely a whisper”“breath caught in”“couldn't help but”“testament to”“unspoken”“Not X, but Y” LLM tic
Computed Slop Score: 99 / 100

Direct Head-to-Head Showdowns & Underlying Model Dossiers

Zero Fluff · Direct Telemetry
Head-to-Head MatchupLab Winner
Ourdream AI vs. Kindroid AI (2026) →Kindroid AI
Ourdream AI vs. Candy AI (2026) →Ourdream AI
Kindroid AI vs. Nomi AI (2026) →Kindroid AI
Ourdream AI vs. Secrets AI (2026) →Ourdream AI
Underlying AI Model DossierWholesale vs Retail
DeepSeek V4.1 vs V3.2 →V4.1 Flash burns 8.2x more t
Z-Image-Turbo (S3-DiT) →11x Retail Markup over Whole
WAI-Illustrious-SDXL →~13x Retail Convenience Mark
Seedream 5.0 + Wan 2.5 →1.8x – 2.4x Markup on Video
Euryale 70B & Hermes 4 405B →Best Value When Bundled in F