asycd

TEV2

TEV2, The End of Generic AI. Your ideas enhanced by live trend awareness and creative memory. Generated instantly on serverless infrastructure.

TEV2 example generation
TEV Research EngineBeta

Query the theme graph

Longer descriptive phrases work best. Click a card to reverse-search the closest themes.

Imagination & intelligent creation

Let AI bring your vision to life — from abstract ideas to vivid, stunning visuals.

Rubber chickens and fruit loops cereal
Pastel prayers landscape

Theme-to-image technology to unlock creativity

Live research

Research logs

Live research into themes, interpretations, news, and source material — the creative context behind each generation.

Streaming · always on

Theme

Theme in focus: “full-screen bloodbath”.

Art Research · Theme intake

A digital landscape overrun by visceral colour, ruined screens, and mythic violence.

Interpretation

Enhanced reading of the theme into a full subject line.

TEV2 creative agent

“Crimson waves crashing over pixelated ruins, steel-clawed dragons weaving through shattered screens, wires sparking in chaotic disarray.”

News

Culture coverage leaning into digital carnage aesthetics and glitch violence metaphors.

Creative press · Visual culture

Across design blogs and culture desks this week, writers keep returning to the same image: the screen as a battlefield. Interfaces bleed, pixels collapse, and mythic creatures crawl through broken UIs. One critic described it as “crimson UI, ruined pixels, dragons in the feed” — less a game reference than a mood for how digital life feels when everything is overstimulated and slightly on fire.

Theme

Nearby theme recovered from creative memory: “bloodbath twist”.

Visual memory graph

Close neighbour to the current theme — shared crimson intensity, different narrative turn.

Key Features

TEV2 converts ideas into expressive visual universes, guided by intelligent research agents and creative memory.

High-Fidelity Generation

  • Photorealistic outputs
  • Art styles & filters
  • Custom resolution settings
  • Prompt refinement tools

Creative Applications

  • Social media content
  • Illustrations & concept art
  • Marketing visuals
  • Product mockups

Seamless Integration

  • Easy export formats
  • Third-party app support
  • API access for developers

Ethical Innovation

  • Uncensored generation
  • No content safety controls
  • Transparent prompt creation
  • Style transformation, not imitation

Documentation

DOCS

TEV2 (Theme Explorer v2) turns simple text themes into vivid art descriptions, then generates images with Stable Diffusion 3.5 or gpt-image. Built by Asycd for canvases, covers, events, and stock visuals — with full ownership of every output.

Why it’s useful

  • Accessibility — no artistic skills required
  • Creativity — abstract themes become visual art
  • Quality — Stable Diffusion 3.5 medium & large
  • Choice — premium gpt-image-2 plus 2.5 Flare or Sunburst (low = FAST; medium/high async)
  • Simplicity — minimal API parameters
  • Ownership — you own all generated images

How the algorithm works

  1. Theme input processing for prompts like “ultimate sadness”
  2. Multi-source enrichment from Reddit, Twitter, Quora, and the web
  3. Agentic processing that streams new visual ideas continuously
  4. Systems theory + combinatorics for description generation
  5. Image generation with Stable Diffusion 3.5 or gpt-image

Endpoints

Stable Diffusion 3.5

POST https://api.asycd.online/generate-image-simple

prompt
Required. Text description of the image you want.
model_size
Optional. medium or large (default medium).
tev_transform
Optional. Defaults true (TEV rewrite first). Set false to send the raw prompt.

gpt-image Premium

POST https://api.asycd.online/generate-image-openai

Quality low is sync and shows the green FAST badge. Higher settings (medium and up, including 2.5 xhigh / max) can return pending + job_id; the app polls GET /generate-image-openai/{job_id}.

prompt
Required. TEV-enhanced server-side unless tev_transform is false.
model
Optional. gpt-image-2, gpt-image-2.5-flare (everyday default), or gpt-image-2.5-sunburst (higher quality / tighter edits). Panel default: gpt-image-2.5-flare. If omitted, the server uses its configured default.
tev_transform
Optional. Defaults true. Panel TEV toggle on = enhance first; off sends false (raw prompt).
quality
Optional. gpt-image-2.5 accepts auto, low (FAST, sync), medium, high, xhigh, or max. gpt-image-2 uses low / medium / high. Panel default: medium.
width / height
Optional. Defaults 1024 × 1024.
output_format
Optional. png, jpeg, or webp.

Python · SD 3.5

import requests

response = requests.post(
    "https://api.asycd.online/generate-image-simple",
    headers={"Authorization": "YOUR_API_TOKEN"},
    data={
        "prompt": "a majestic dragon in a fantasy landscape",
        "model_size": "large",
        "tev_transform": True
    }
)

curl · gpt-image

curl -sS -X POST \
  "https://api.asycd.online/generate-image-openai" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A luminous jellyfish in deep ocean",
    "width": 1024,
    "height": 1024,
    "quality": "low",
    "output_format": "png",
    "model": "gpt-image-2.5-flare",
    "tev_transform": true
  }'

Response fields

  • success — generation status
  • image_base64 — image as base64
  • filename — generated file name
  • generation_time — seconds
  • error — message or null

SD 3.5 Medium / Large

Medium ~15–45s; Large ~30–90s for higher detail and print-ready work.

gpt-image

Premium OpenAI path. Choose gpt-image-2, 2.5 Flare (everyday default), or 2.5 Sunburst (slower, tighter). Only low is FAST (sync). Higher qualities, including 2.5 xhigh / max, use async job polling and take longer.

Full reference on asycd.online

Access

Get access to the API

Your TEV2 API token unlocks unlimited AI image generation. After you register, we review requests within 12–24 hours and email your unique token on approval.

Include the token in the Authorization header when calling /generate-image-simple or /generate-image-openai. Send a prompt (and optional knobs) — images typically return in 15–90 seconds. No rate limits. You own every image.

Start creating immediately after your token arrives.

  1. 01Submit your registration
  2. 02Wait for approval (12–24 hours)
  3. 03Generate with your API token
Request access