Faymas: What It Is, What Its Prompts Do, and Which Ones Actually Work

Faymas is an AI prompt marketplace and creative studio that operates across multiple web properties — and depending on which one you land on, you'll get a very different picture of what the platform actually is. One property is a full image and video generation studio with subscription tiers and nine integrated AI models. Another is a prompt marketplace where individual creators publish prompts and earn money each time someone copies them. A third republishes prompts but attributes everything to a single admin account, not to the creators who originally wrote them. Most English pages that mention Faymas describe only one of these faces.

The two sides of Faymas

The generation studio side lets you pick from models including Nano Banana, Veo, Wan, GPT Image 2, Kling, Seedream, Seedance, HappyHorse 1.1, and Grok Imagine. You choose text-to-image or image-to-image, set an aspect ratio (1:1, 3:4, 2:3, 3:2, 9:16, 16:9), pick a resolution (1K, 2K, 4K), and the platform shows you a credit estimate before you generate. Pricing runs from $9.9/month for 220 credits up to $39.9/month for 2,000 credits, with a yearly discount of up to 23% and a pay-as-you-go option. Every plan includes access to all models, all prompts, a 7-day generation history, flexible aspect ratios, and ad-free browsing.

The marketplace side is a different animal. Here, creators publish prompts — mostly tagged for two consumer-facing AI models rather than the nine studio models — alongside the actual AI output each prompt produced. Every prompt page shows three engagement metrics: total likes, total copies, and total saves. The monetization model is built around one action: the copy. When a user copies a prompt to paste into their own AI tool, the original creator earns. The platform explicitly states that income is not capped by follower count or join date — it's driven by whether prompts are genuinely useful.

What the prompts actually look like

Most prompts on the marketplace follow a consistent structural pattern. They run long — 150 to 300 words — and are heavily weighted toward identity preservation. A representative example reads:

Ultra-realistic cinematic 8K vertical mirror selfie portrait of an elegant South Asian woman wearing a rich deep red silk saree with intricate antique-gold woven floral motifs and a broad traditional golden border. She is wearing a matching deep red short-sleeve blouse with subtle golden embroidery on the sleeves. The saree is draped gracefully in a classic Indian style with natural pleats and realistic fabric folds. Her hair is styled in an elegant traditional low bun adorned with fresh white jasmine flowers, with a few soft face-framing strands falling naturally near her cheeks. She wears delicate traditional gold jhumka earrings, a fine gold necklace with a small pendant, a subtle waist chain, golden bangles, and a tiny elegant bindi. Her hands feature detailed traditional mehndi/henna designs and neatly manicured deep-red nails. She is taking a mirror selfie with a premium rose-gold smartphone held naturally in one hand, partially covering the side of her face. Soft confident expression with a gentle natural smile, realistic facial proportions, expressive eyes, natural skin texture, subtle makeup, soft red lipstick. Background: luxurious modern bedroom with large floor-to-ceiling mirrored wardrobe panels, warm ambient ceiling lights, elegant curtains, a neatly arranged bed, polished light-colored flooring, and realistic reflections showing the back of the saree and hairstyle. Photography: ultra-realistic DSLR quality, cinematic warm lighting, natural shadows, realistic skin pores, highly detailed saree texture, authentic jewelry reflections, accurate mirror reflection, shallow depth of field, professional fashion photography, HDR, photorealistic, 8K, vertical 9:16 composition. Identity preservation: Use the uploaded reference image only as the face/identity reference. Preserve the exact facial structure, eyes, eyebrows, nose, lips, jawline, skin tone, and distinctive facial features. Do not alter the person's identity or facial proportions.

A shorter, more atmospheric example:

Keep my exact facial features, identity, hairstyle, skin tone, face shape, and appearance unchanged. Transform me into a cinematic dark gothic character standing alone in a vast open field. Long black trench coat flowing in the wind. Large flock of black crows flying around me at different distances. Stormy overcast sky. Moody atmosphere. Black and white photography. Natural wind movement in hair and clothing. Looking away from camera. High contrast cinematic lighting. Professional fashion editorial photography. Ultra realistic. Photorealistic. Masterpiece composition.

Both share a core mechanic: upload a reference photo, lock the subject's identity, then transform the scene, clothing, or mood around that fixed identity. This “identity lock” pattern is the dominant prompt architecture on the platform — the majority of visible prompts are portrait photography prompts built around face preservation.

What the engagement numbers actually show

The marketplace's blog claims 40,000+ users and 125,000+ total prompt copies since public beta launched in July 2026. But the visible engagement on individual prompts tells a more granular story. The most-copied prompt visible on the marketplace homepage at the time of research had 134 copies and 24 saves. Several prompts showed 48 to 56 copies. Many had zero likes, zero copies, and zero saves. The distribution is steep — a small number of prompts drive the bulk of copy activity, while a long tail sits at or near zero.

This matters because the entire earnings model is tied to copies. If you're a creator evaluating whether to invest time publishing here, the question isn't just “can I write good prompts” — it's whether the platform's discovery and search surface enough volume for your work to get found.

The copy-based monetization model — what's stated and what's missing

The platform's founder describes the vision as becoming “the monetization layer for AI creators, helping turn AI knowledge into digital assets.” The creator onboarding guide spells out a five-step process: apply through the Founding Creator Program, choose a niche, write prompts that get copied, publish consistently, and track performance via likes/copies/saves metrics.

What's stated clearly: earnings are per-copy, not per-follow or per-like. You don't need an existing audience. The platform is free to browse and copy. Creators are onboarded through an application-based program, not open registration.

What's not stated anywhere on the platform: the per-copy payout rate, minimum payout thresholds, payment methods, or what percentage of revenue goes to the platform versus the creator. The blog says creators “earn” but never specifies how much a single copy is worth in actual money. This is the single biggest gap in the platform's public documentation. Whether the economics work for creators at current copy volumes — where a top prompt has 134 copies — depends entirely on a rate the platform doesn't disclose.

Prompt categories and what's actually available

The generation studio's prompt library organizes prompts into topic collections with visible counts:

That's roughly 700+ prompts across the studio's library. The marketplace side lists prompts individually with model tags (mostly “chatgpt” and “gemini”) rather than by model-specific collections.

Which prompts actually work

Based on the engagement data visible on the marketplace, three prompt types consistently outperform others:

  1. Identity-preserving portrait prompts with reference images — prompts that instruct the AI to lock facial features from an uploaded photo, then build a new scene around them. These dominate the top-copied list.
  2. Cultural and festival-specific prompts — prompts tied to specific cultural contexts (traditional attire, devotional scenes, festival celebrations) get copied at higher rates than generic portrait prompts, likely because they solve a specific use case rather than offering a general aesthetic.
  3. Couple and relationship portrait prompts — split-frame compositions, long-distance couple portraits, and romantic selfie prompts consistently appear in the top-copied tier.

What doesn't get copied: abstract art prompts, open-ended concept prompts without reference image requirements, and prompts without visible example outputs. The platform's own creator guide confirms this pattern — it explicitly states “a prompt paired with a strong example output performs far better than plain text alone.”

The three-property problem

Here's what no single English page about Faymas explains: the platform operates across at least three distinct web properties, and they don't all behave the same way.

The generation studio is a paid tool — you subscribe, get credits, and generate. Its prompt library is a curated collection organized by topic, with no per-creator attribution visible on the library listing pages. It's a product.

The marketplace is a free-to-browse community platform where individual creators have profiles, publish prompts, and earn per copy. Engagement metrics (likes, copies, saves) are visible on every prompt card. It's a social platform.

The third property republishes prompts from the marketplace but attributes them all to a single admin account rather than to the original creators. It has no visible engagement metrics, no creator profiles, and no monetization features. It reads like a content aggregation layer — possibly for SEO, possibly as a mirror — but it breaks the creator attribution model that the marketplace is built on. [This characterization of the third property's purpose is an inference, not a stated fact on any page.]

If you're evaluating Faymas as a creator, you need to know which property you're on. Publishing on the aggregation property would earn you nothing, because it doesn't track copies to individual creators. [Inference based on the absence of creator attribution and engagement metrics on that property.] Publishing on the marketplace is where the monetization mechanism lives. Using the generation studio is where you'd actually run the prompts.

How to actually use Faymas prompts

The platform's own documentation lays out a six-step workflow for the marketplace:

  1. Browse or search the prompt library
  2. Preview the AI-generated output attached to each prompt
  3. Tap “Copy” on the prompt you want
  4. Paste it into your AI tool of choice (ChatGPT, Gemini, Midjourney, etc.)
  5. Upload a reference photo if the prompt requires identity lock
  6. Generate your result

For the generation studio, the workflow is different — you stay within the platform: choose a model, write or paste a prompt, set aspect ratio and resolution, see the credit cost, generate, and download. No external tool needed.

What Faymas gets right and where it's thin

What works: the identity-lock prompt pattern is genuinely useful. If you've ever tried to generate consistent portraits of the same person across different scenes in ChatGPT or Gemini, you know it's hard to maintain facial consistency. The prompts on Faymas encode a specific, tested structure for achieving this — “Use the uploaded image as the ONLY source of the subject's identity and facial appearance. Preserve the user's exact recognizable identity with maximum fidelity: facial structure and proportions, face shape…” — and that structure works. Copying one of these prompts and adapting the scene description is faster than engineering the identity-preservation language from scratch.

What's thin: the platform is overwhelmingly focused on portrait photography, particularly South Asian cultural contexts. If you're looking for prompts in other domains — coding, writing, business analysis, data visualization — there's very little. The model coverage is also uneven. The marketplace tags prompts as “chatgpt” or “gemini” but rarely as Midjourney, FLUX, Veo, Runway, or Kling — despite the platform claiming to support all of them. The generation studio side does integrate those models, but the marketplace prompts are overwhelmingly built for two.

What's missing: the per-copy earnings rate, any public revenue data, and any creator testimonials with actual income numbers. The platform's launch announcement on a professional networking site had 5 likes and 1 comment at the time of research — early traction, but not validation at scale.

Bottom line

Faymas is two products wearing one name: a capable multi-model generation studio and a prompt marketplace with a copy-based creator economy. The prompts that work best are identity-preserving portrait prompts that solve a specific problem — maintaining facial consistency across scenes. The prompts that don't work are generic, reference-free, or lack visible example outputs. The monetization model is promising in theory but unproven in practice, because the platform doesn't disclose what a copy is actually worth.