Create yourself. Invent anyone. Build whole worlds.
Siren handles identity training, instant character creation, world-building, scene composition and post-production — so you can focus on the story, not the toolchain.
Invent a character in seconds — no training
Siren has two ways to create an identity. Create Me trains on your photos for the highest likeness. Create a Character invents someone brand new from a prompt or a few taps, and has them ready to create with right away.
- Describe Them — write who you're imagining in plain words
- Build Them — guided options for gender, age, body type, hair, eyes, style and more
- Surprise Me — Siren invents the look and the concept, with an optional theme
- Realistic, Anime or Fantasy art styles
- Pick from four generated faces; Siren builds a consistent reference set from your choice
What they look like — and who they are
Every character has a visual identity and an optional character profile. Go deep for story characters, or skip it entirely: a brand model for your candle company needs a name, not a childhood.
- Name, age, gender, occupation and appearance
- Personality, background, relationships, speaking style and scenario
- Public description and tags for Explore
- Private creator instructions that are never shared
- Attach worlds so the character's scenes stay on-world
Worlds: generative world packs
A World (or lorebook) holds written lore and visual information — so it informs what Siren generates, not just what it writes.
- Characters, locations, events and aesthetics in one place
- Use This World — choose who's in the scene, then pick a location and event
- Bring your own characters into worlds other creators publish
- One world feeds images, stories and video
Vale University
Create with any character
Pick a character — yours, or a community one you're allowed to use — choose Image, Story or Video, and set the scene. Pull the location and event straight from a world.
- Image, Story (multi-frame) and Video from one screen
- Save, add to project, make variations or animate the result
- Instant mode matches the face from the character's reference photos
Publish characters and worlds
Everything is private by default. Publish when you're ready, and choose exactly what goes out and what others can do with it. Publishing is free.
- Private (only you), Unlisted (anyone with the link) or Public (in Explore)
- Choose what to publish: identity images, profile, linked world, sample scenes
- Use allowed — others create with it while it stays yours
- Remix allowed — others make their own version, credited to you
- Optional "no commercial use" label
- Creator profiles, likes, saves and use counts; browse Characters and Worlds in Explore
Chat on the web and by SMS
Characters talk the way you wrote them — personality, background, speaking style and worlds all shape the reply. Chat in the app, or give a character a phone number so anyone who texts it chats with them by SMS.
- Regular characters reply instantly on every plan
- Replies can include a photo of the character (MMS on SMS)
- Chat with community characters when their creator allows use
- SMS chat is on every plan
- 18+ character chat — Ultimate. 18+ characters are private or unlisted only — never public — and are answered on Siren's own GPUs
✨ Lock This Identity
Characters work the moment you create them. When you want maximum consistency across future generations, Lock This Identity trains on the character's reference images — the same premium, high-fidelity training behind Create Me. Training becomes an upgrade, not an entrance fee.
- Optional, per character, whenever you're ready
- Generate extra reference images first — more images give a better lock
- Create Me and locked characters share the same trained-identity pipeline
Images, video & trained identities
Ultra realistic AI images of you, on demand
Siren's four-engine pipeline — Studio (FLUX.1), Lifestyle (SDXL), Pro (FLUX.2), and Z-Image Turbo — renders photorealistic portraits that hold your likeness across every scene, outfit, and lighting setup.
- Per-profile identity LoRAs trained on your reference photos
- 5-slot stack: face, body, anatomy, intimate, detail
- Studio-grade quality tier for magazine-ready output
AI videos and reels from a still
Animate any Siren portrait into short-form video — perfect for reels, TikTok loops, and product spots. Pose, motion, and camera moves are prompt-controlled.
- Image-to-video pipeline with identity-locked frames
- Reel-ready 9:16, 1:1, and 16:9 outputs
- Batch generation for content calendars
Make AI content of yourself in minutes
This is the Create Me path. Upload three reference photos. Siren trains on your face and body, then drops you into any setting you describe — beach, studio, magazine cover, anywhere.
- Three-photo identity bootstrap, optional expansion later
- 167-prompt curated gallery to start fast
- Private by default — your photos and outputs stay yours
Create a virtual influencer from scratch
Design a fully synthetic persona — face, body, style — and build a consistent content library around them. Start instantly with Create a Character, then Lock This Identity so your brand stays on-model across thousands of generations.
- Start from a prompt, guided builder or Surprise Me — no photos needed
- Synthesize identity from a curated reference set
- Style and look-mode presets locked per persona
- Full output history and asset library per profile
