Scrya — AI Companions & Character Chat

Build cinematic prompts visually, inject them into https://grok.com with one click, chat with AI characters powered by your own LLM, and auto-capture everything you generate to your own cloud storage. --- 🎨 Build prompts like a director - 🧱 Visual prompt pipeline — Character → Outfit → Scene → Action → final prompt → one-click inject straight into Grok. - 📚 Thousands of curated prompts across 13 tabs: actions, characters, outfits, lighting, effects, locations, camera angles, dialogue. - 🛍️ Prompt-pack marketplace — subscribe to community + official packs from Scrya; browse by category in a built-in pack browser. - 👩‍🎨 Sims-style character builder — 25 attributes across identity, ethnicity, hair, face, expression, style. Live preview, cloud save. 🤖 Chat with your own characters - 🎭 Victoria & custom chatbots — built-in flirtatious chat companion plus your own personas. Each one carries its own system prompt, voice, and visual identity. - 🎬 Method-acting roleplay from a favorite — click any favorite image to instantly become that character's chat partner. The model adopts voice, mannerisms, mini-lore over multiple turns. - ✨ AI persona generator — type "Axel Foley" + source "Beverly Hills Cop" → your LLM expands it into a rich method-acting brief (voice, backstory, catchphrases, scene-entry style). Edit before you save. - 🎞️ Import any character from IMDb — pick a movie that's been indexed, tap Add on Axel, Vivian, John Wick — they arrive as a fully-briefed chatbot with the movie poster as avatar. - 🌐 Japanese support from day 1 — flip EN | 日本語 in the chat header; system prompts ship as Dotprompt files for easy translation extension. 🔌 LLM your way — local, hosted, or BYO - 🆓 Bring your own local LLM — Ollama, LM Studio, llama.cpp, vLLM, any OpenAI-compatible server. Fully private, zero cost. Free LFM 2.5 link for users with no model yet. - ☁️ Hosted fallback — OpenAI, Anthropic, Gemini, xAI Grok, DiffusionGemma — plug in your API key and go. - 🔐 Your keys, your machine — keys live in extension storage; nothing hits our servers. 🎙️ Voice & TTS - 🗣️ ElevenLabs voice picker — full-screen browser with previews, filter by language/gender/accent, one-tap to assign to any character. - 🎚️ Multi-backend TTS — Kokoro — point at your own endpoint. ☁️ Your own cloud archive - 🔍 Auto-capture from Grok — every image and video you generate is synced to Supabase + Cloudflare R2 under your account. No platform lock-in. - 🗂️ Cloud lists + trigger words — tag favorites into curated lists for LTX-2 training, export with one click. - 🖼️ R2 CDN thumbnails — every prompt gets a 256×256 WebP thumbnail, served globally. 🚀 Get started in 60 seconds 1. Install — open the sidepanel on grok.com. 2. Sign in with Google or X. 3. Pick a tab, build a prompt, hit Inject. 4. (Optional) Wire your LLM key in Settings → chat with Victoria. --- Free to install. Cloud sync requires a free Scrya account. Bring your own LLM keys for chat.