P-Image-Ideogram

P-Image-Ideogram is Pruna AI’s text-to-image model, built in collaboration with Ideogram. It offers four thinking levels that trade off speed, price, and quality: Very Low, Low, Medium, and High.

P-Image-Ideogram fashion lookbook JSON (Medium)

Use it for photoreal campaigns, product and people-led visuals, precise typography, and structured JSON prompts with hex colors and bounding boxes. Pair Medium for everyday production, Very Low or Low for drafts and high-volume exploration, and High at 2K or above when prompts are detailed or text rendering must be crisp.

At a glance

Metric	P-Image-Ideogram
Speed (1K, prompt upsampling on)	~0.6 s (Very Low) to ~7.5 s (High), depending on `thinking`, `image_size`, and `prompt_upsampling`
Price	~$0.003 (Very Low) to ~$0.03 (High) per image
Thinking levels	`very_low`, `low`, `medium` (default), `high`
Text rendering	Strong; use High at 2K+ for difficult typography
API model header	`Model: p-image-ideogram`

Thinking modes

Speeds below are measured at 1K with prompt_upsampling: true. Higher thinking values generally work best with more detailed prompts. On vague prompts, High can add unrequested detail. Low is often better for simple, non-specific prompts.

`thinking`	Cost	Speed	Best for
`very_low`	~$0.003	~1.6 s	High-volume generation, drafts, exploration, A/B testing
`low`	~$0.0075	~1.8 s	Cost-efficient production at scale; simple prompts
`medium`	~$0.015	~2.4 s	Everyday production: best default balance
`high`	~$0.03	~3.6 s	Complex prompts, fine detail, brand-critical assets, difficult text at 2K+

Strengths and limits

Strengths	Details
Photorealism	Very strong performance on people, products, and lifestyle scenes
Text rendering	Reliable on-image typography when you use `high` at 2K+
Flexible prompting	Natural language and structured JSON; hex colors and bounding boxes supported
Speed and cost	Fast and affordable across all thinking levels

Limits	Impact
Creative styles	Non-photoreal and stylized looks are still improving
Low-mode artifacts	`low` can show more artifacts than `medium` or `high`
Text at low thinking	Difficult text rendering does not work well in `very_low` or `low`. Use `high` at 2K
Mode choice	Higher thinking is not always better: match the mode to the prompt. `low` for simple prompts, `high` for complex ones. On vague prompts, `high` may add unwanted detail

Feature highlights

Four capabilities that define how to use P-Image-Ideogram in production. Open Show prompt under each example to copy the text.

Four generation levels via thinking: different quality / latency / cost.

Note

Higher modes work best with detailed prompts. On a vague prompt, high can invent unrequested detail. Prefer low or medium for simple exploration.

Street-style editorial

fashion_street_style Very Low — **Very Low**

fashion_street_style Medium — **Medium**

Show prompt

Editorial fashion photograph of a fictional young creative professional walking through a rainy residential neighborhood in Copenhagen, oversized neutral tailoring, silver accessories, natural walking movement, soft overcast light, wet pavement reflections, authentic street-style photography, realistic face and hands, no visible logos, no text.

image_size sets the output resolution budget (default "1K"). Use "1K" for faster drafts when typography is not critical. Prefer 2K with high for brand-critical or typography-heavy work.

Beauty shade at thinking: high. Drag to compare typography and skin detail:

↔ Drag 1K 2K

Show prompt

Vertical beauty campaign for a fictional lipstick shade.

Editorial close-up of a fictional model with luminous skin, soft gaze, dark hair pulled back, and bold red lipstick. Clean studio backdrop, shallow depth of field, realistic skin texture. Negative space on the left for typography.

Exact visible text only:
Product name exactly: "ROUGE ATELIER"
Shade line exactly: "No.07 - Riviera"
No logos, no barcodes, no additional text.

CRITICAL TYPOGRAPHY: perfectly formed and correctly spelled letters; no overlapping glyphs, duplicated syllables, missing or extra characters.

Typography is a strength, but difficult text does not work in very_low or low. Use ``high`` at 2K, keep copy short, and use an Exact visible text only: + CRITICAL TYPOGRAPHY: block in the prompt.

Beauty shade campaign

Show prompt

Vertical beauty campaign for a fictional lipstick shade.

Editorial close-up of a fictional model with luminous skin, soft gaze, dark hair pulled back, and bold red lipstick. Clean studio backdrop, shallow depth of field, realistic skin texture. Negative space on the left for typography.

Exact visible text only:
Product name exactly: "ROUGE ATELIER"
Shade line exactly: "No.07 - Riviera"
No logos, no barcodes, no additional text.

CRITICAL TYPOGRAPHY: perfectly formed and correctly spelled letters; no overlapping glyphs, duplicated syllables, missing or extra characters.

Pass a JSON string as prompt for compositional control: hex colors, bounding boxes, and typed elements (text / obj). Set prompt_upsampling: false when the spec is final.

JSON is worth it when you need exact layout, brand hex colors, or repeatable composition — posters, labels, paid social, slide decks. For quick mood exploration, natural language is usually enough.

P-Image-Ideogram uses the same Ideogram 4.0 structured-caption schema documented in JSON Prompting (Ideogram 4.0). The three top-level fields:

high_level_description — one- or two-sentence summary of the whole image (strongly recommended)
style_description — optional aesthetics, lighting, and medium; use photo or art_style, not both; optional color_palette of uppercase #RRGGBB hex codes (up to 16 for the image)
compositional_deconstruction — required; background first, then elements[]

Each element is "type": "obj" (a subject or object, with desc) or "type": "text" (literal copy in text plus desc). Optional bbox places an element with normalized coords [y_min, x_min, y_max, x_max] on a 0–1000 grid (origin top-left); omit bbox to let the model place freely. Per-element color_palette accepts up to five hex codes.

Note

Key order matters. The model was trained on captions with a consistent field order — follow the schema in the Ideogram guide rather than reordering keys. See Configuration below for a full API request example.

Lookbook cover with JSON placement for the masthead and collection lines

fashion_lookbook_json Very Low — **Very Low**

fashion_lookbook_json Medium — **Medium**

Show prompt

{
  "aspect_ratio": "3:4",
  "high_level_description": "A luxury fashion lookbook cover for a fictional ready-to-wear house. Minimal concrete gallery, sculptural ivory coat on a fictional model, bold masthead and season line. Prioritize large crisp typography. CRITICAL TYPOGRAPHY: Every visible letter must be perfectly formed and correctly spelled. No overlapping glyphs, no duplicated syllables, no missing/extra characters, no garbled noise, no random digits inside words, and no speed/timing/performance badges or stopwatch callouts. Prefer fewer, larger text elements over dense copy.",
  "compositional_deconstruction": {
    "background": "Minimalist light-gray concrete gallery with soft side lighting and long elegant shadows.",
    "global_text_rules": "All visible text must be sharp and correctly spelled. No tiny captions. No extra logos.",
    "elements": [
      {
        "type": "text",
        "bbox": [
          70,
          80,
          140,
          920
        ],
        "text": "ATELIER NOIR",
        "desc": "Small widely tracked uppercase black masthead across the top."
      },
      {
        "type": "obj",
        "bbox": [
          160,
          180,
          820,
          820
        ],
        "desc": "Fictional model in sculptural ivory coat, dramatic side light, premium fashion photography, realistic face and hands."
      },
      {
        "type": "text",
        "bbox": [
          860,
          80,
          940,
          920
        ],
        "text": "THE NEW ESSENTIALS",
        "desc": "Large bold black headline across the lower third."
      },
      {
        "type": "text",
        "bbox": [
          950,
          80,
          990,
          920
        ],
        "text": "FALL COLLECTION",
        "desc": "Medium secondary line under the headline."
      },
      {
        "type": "text",
        "bbox": [
          200,
          80,
          260,
          400
        ],
        "text": "Look 01",
        "desc": "Small elegant gray index label."
      }
    ]
  }
}

Domain use cases

Examples from the partner kit, grouped by vertical. Each row shows Very Low, Low, Medium, and High. Open Show prompt to copy the text.

Useful for lookbook moods, street-style editorials, show collateral, and colorway lifestyle strips before a shoot.

Colorway lifestyle strip

fashion_colorway Very Low — **Very Low**

Show prompt

Wide ecommerce product lifestyle strip for a fictional sneaker colorway launch.

A single generic white sneaker on a sunlit concrete step, photoreal materials, no logos on the shoe. Clean negative space on the right.

Exact visible text only:
Product line exactly: "AERO RUN"
Colorway exactly: "Salt / #E8E2D6"
No other text, no logos, no watermark.

Runway invitation

fashion_runway_invitation Medium — **Medium**

fashion_runway_invitation High — **High**

Show prompt

Vertical fashion-week invitation card on warm ivory paper with a minimal architectural photograph of a concrete gallery doorway in the upper half.

Exact visible text only:
Title exactly: "ATELIER NOIR"
Line 2 exactly: "FALL 26 SHOW"
Line 3 exactly: "19 SEPT - PARIS"
No logos, no QR codes, no additional text, no watermark.

On-model campaign moods, shade storytelling, and social beauty atmospheres. These are creatives, not brand packshot replacements.

Beauty ritual mood

Show prompt

Vertical beauty campaign mood for a fictional skincare ritual.

Editorial close-up of a fictional young woman with luminous skin, soft natural makeup, morning bathroom light through frosted glass. Towel-wrapped hair, calm expression, shallow depth of field, realistic skin texture. No product bottle in sharp focus: atmosphere and people first. Clean negative space on the left for typography.

Exact visible text only:
Line 1 exactly: "MORNING GLOW"
Line 2 exactly: "Ritual 01"
No logos, no barcodes, no additional text.

Performance ads, OOH, and UGC-style people shots when you need readable headlines and fast campaign iteration.

OOH billboard

Prefer medium or high at 2K when the typography is large.

marketing_ooh_billboard Medium — **Medium**

Show prompt

Vertical out-of-home billboard mockup for a fictional lifestyle brand. Photoreal city dusk street scene with a large illuminated billboard. Exact visible text only on the board: headline "OWN THE NIGHT" and subline "CITY LIGHTS COLLECTION". No logos, no watermarks, no extra text.

Paid social (JSON layout)

marketing_paid_social_json Very Low — **Very Low**

marketing_paid_social_json Low — **Low**

marketing_paid_social_json Medium — **Medium**

marketing_paid_social_json High — **High**

Show prompt

{
  "aspect_ratio": "4:3",
  "high_level_description": "A polished vertical paid-social ad for a fictional sparkling drink brand called LUMI. Bright sunlit product hero, bold headline, short subcopy, and a clear CTA pill. Brand accent color exactly #FF5A1F. Prioritize large, crisp, highly legible typography. CRITICAL TYPOGRAPHY: Every visible letter must be perfectly formed and correctly spelled. No overlapping glyphs, no duplicated syllables, no missing/extra characters, no garbled noise, no random digits inside words, and no speed/timing/performance badges or stopwatch callouts. Prefer fewer, larger text elements over dense copy.",
  "compositional_deconstruction": {
    "background": "Soft cream-to-peach gradient with subtle sun flare in the upper right. Clean commercial advertising look.",
    "global_text_rules": "All visible text must be sharp, correctly spelled, and easy to read on mobile. Use fewer words and larger font sizes. Do not generate tiny legal copy. CRITICAL TYPOGRAPHY: Every visible letter must be perfectly formed and correctly spelled. No overlapping glyphs, no duplicated syllables, no missing/extra characters, no garbled noise, no random digits inside words, and no speed/timing/performance badges or stopwatch callouts. Prefer fewer, larger text elements over dense copy.",
    "elements": [
      {
        "type": "text",
        "bbox": [
          60,
          70,
          120,
          400
        ],
        "text": "LUMI",
        "desc": "Small bold brand wordmark in vivid #FF5A1F at top-left."
      },
      {
        "type": "obj",
        "bbox": [
          140,
          220,
          680,
          780
        ],
        "desc": "Tall glass bottle of pale citrus sparkling drink with condensation, centered, photoreal, blank frosted label panel."
      },
      {
        "type": "text",
        "bbox": [
          700,
          80,
          860,
          920
        ],
        "text": "Stay\nlight.",
        "desc": "Huge bold near-black headline stacked on two short lines."
      },
      {
        "type": "text",
        "bbox": [
          880,
          80,
          940,
          700
        ],
        "text": "Zero sugar. Full sparkle.",
        "desc": "Medium dark-gray subcopy in one large line."
      },
      {
        "type": "obj",
        "bbox": [
          880,
          760,
          960,
          940
        ],
        "desc": "Rounded CTA pill filled with exact brand color #FF5A1F."
      },
      {
        "type": "text",
        "bbox": [
          905,
          790,
          945,
          920
        ],
        "text": "Shop now",
        "desc": "Bold white CTA label centered in the pill."
      }
    ]
  }
}

Key art, title banners, and pitch moods when you need cinematic stills with clean title typography.

Game key art banner

Show prompt

Cinematic game key art banner for a fictional action RPG. Lone armored knight on a cliff overlooking a burning medieval city at dusk, volumetric fog, embers, dramatic orange sky, ultra-detailed armor, shallow depth of field. Clean negative space for title typography.

Exact visible text only:
Title exactly: "EMBER FALL"
Subtitle exactly: "ACT I - THE LAST GATE"
No logos, no watermarks, no additional text.

Delivery heroes, recipe banners, and dining lifestyle shots for campaigns and social.

Delivery lifestyle banner

Show prompt

High-conversion food delivery website hero banner. Sunny urban terrace where friends have just received a large food delivery: open kraft bags, unbranded takeaway containers, fresh pasta and salads, spontaneous sharing. Bold left panel with headline exactly "LUNCH JUST LANDED" in large white sans-serif on delivery green #06C167. Photoreal, joyful, no brand logos, no watermarks, no extra text.

Recipe banner

Show prompt

Recipe website hero banner. Bright contemporary home kitchen, late afternoon. Home cook tossing tomato-basil pasta; fresh ingredients on the counter; plated bowl beside the pan. Soft natural window light. Headline exactly "COOK IT TONIGHT" in large bold white sans-serif on the left. No logos, no brands, no extra text.

Lifestyle and night-drive scenes where the car sits in context with people, not as a sterile packshot.

Weekend coastal lifestyle

automotive_weekend Very Low — **Very Low**

Show prompt

Cinematic lifestyle photograph of a couple loading weekend bags into a generic compact car at a coastal overlook.

Vehicle is secondary: people, wind, ocean light, and travel mood first. No readable manufacturer badges or license plates. Photoreal golden-hour light.

Exact visible text only:
Headline exactly: "WEEKEND OUT"
No logos, no dealer chrome, no watermarks.

Night city route (JSON)

Denser specs hold up better at medium or high.

automotive_night Very Low — **Very Low**

Show prompt

Premium automotive hero. Rainy city at night, neon reflections on wet asphalt. Generic dark hatchback, no brand marks or license plate. Stylish driver in cinematic close-up through the side window. Headline exactly "NIGHT ROUTE", vehicle name "URBAN GT". Photoreal, no logos, no watermarks.

Destination cards, booking visuals, and family travel heroes for apps and seasonal campaigns.

Itinerary day card

travel_itinerary Very Low — **Very Low**

Show prompt

Vertical travel-app itinerary card.

Upper half: photoreal dusk view of trams on a steep Lisbon street, warm lights, no people in the foreground. Lower half: clean cream panel.

Exact visible text only:
Day label exactly: "DAY 03"
City exactly: "LISBON"
Activity exactly: "Miradouro sunset walk"
No logos, no maps UI chrome, no additional text.

Global travel hero

Show prompt

Global travel campaign showing three generations of one family arriving by a small wooden boat at a peaceful Greek island harbor. Two grandparents, two parents and two children interact naturally as they step onto the dock. White village buildings rise behind them, sparkling morning sea, realistic faces, hands, clothing and family relationships, premium cinematic travel photography. Headline exactly: "THE STORIES WE BRING HOME". No additional text, no logos, no duplicated people.

Room atmospheres and lifestyle staging. Brands still supply their own furniture SKUs; these shots set mood and context.

Marketplace lifestyle staging

furniture_lifestyle Very Low — **Very Low**

Show prompt

Premium furniture marketplace hero banner. Bright contemporary urban living room, Sunday morning. Generic modular sand sofa with coordinated oak coffee table, cream lounge chair, walnut media unit. Young couple relaxing naturally. Soft diffused daylight. Headline exactly "ROOM TO LIVE". No logos, no brand names, no price tags.

Designer concept moodboard

furniture_moodboard Very Low — **Very Low**

Show prompt

Premium 4:3 interior design concept board for a boutique hotel bedroom inspired by quiet Mediterranean modernism. Dominant photoreal bedroom (~65%) plus material samples and color swatches in an editorial column. Exact visible text only: "QUIET MEDITERRANEAN". No logos, no prices, no supplier names.

Lesson covers, language-app art, and launch slides with clear on-image typography.

Immersive history lesson

education_history Very Low — **Very Low**

Show prompt

Educational lesson cover. Cinematic historical reconstruction of a busy printing workshop in Renaissance Venice around 1500. Craftspeople operating an early wooden printing press. Warm workshop light. Headline exactly "HOW PRINT CHANGED THE WORLD". Photoreal, educational, no logos, no modern objects.

Language learning app

education_language Very Low — **Very Low**

Show prompt

Polished 3:4 mobile illustration for a language-learning app. Cheerful young learner at a lively outdoor market in Barcelona ordering fruit in Spanish. Playful flat 2.5D style. Exact visible text only: "REAL-WORLD SPANISH" in bold rounded sans-serif on a green banner. No logos, no photorealism.

Product launch slide

education_slide_json Very Low — **Very Low**

education_slide_json Medium — **Medium**

Show prompt

{
  "aspect_ratio": "16:9",
  "high_level_description": "A polished dark SaaS product launch slide for P-Image-Ideogram. Navy-to-purple gradient with a soft grid. Title and badge at top left, four vertical generation-level cards across the center, footer trade-off bar at the bottom. Prioritize large, crisp typography.",
  "compositional_deconstruction": {
    "background": "Dark navy-to-deep-purple gradient with a subtle geometric grid and soft purple glows.",
    "global_text_rules": "All visible text sharp and correctly spelled. Prefer fewer, larger text elements.",
    "elements": [
      {
        "type": "text",
        "bbox": [
          60,
          50,
          120,
          420
        ],
        "text": "P-Image-Ideogram",
        "desc": "Large bold white title top-left."
      },
      {
        "type": "text",
        "bbox": [
          70,
          450,
          120,
          700
        ],
        "text": "4 GENERATION LEVELS",
        "desc": "Small purple pill badge next to the title."
      },
      {
        "type": "text",
        "bbox": [
          140,
          50,
          190,
          700
        ],
        "text": "Pruna AI \u00d7 Ideogram \u00b7 text-to-image model",
        "desc": "Small light-gray subtitle under the title."
      },
      {
        "type": "obj",
        "bbox": [
          240,
          60,
          820,
          260
        ],
        "desc": "Card VERY LOW: dark rounded panel with label, ~1.6s at 1K, $0.003/image."
      },
      {
        "type": "text",
        "bbox": [
          260,
          90,
          320,
          240
        ],
        "text": "VERY LOW",
        "desc": "Card title."
      },
      {
        "type": "text",
        "bbox": [
          340,
          90,
          500,
          240
        ],
        "text": "~1.6s at 1K\\n$0.003/image",
        "desc": "Speed and price lines."
      },
      {
        "type": "obj",
        "bbox": [
          240,
          280,
          820,
          480
        ],
        "desc": "Card LOW: dark rounded panel."
      },
      {
        "type": "text",
        "bbox": [
          260,
          310,
          320,
          460
        ],
        "text": "LOW",
        "desc": "Card title."
      },
      {
        "type": "text",
        "bbox": [
          340,
          310,
          500,
          460
        ],
        "text": "~1.8s at 1K\\n$0.0075/image",
        "desc": "Speed and price lines."
      },
      {
        "type": "obj",
        "bbox": [
          240,
          500,
          820,
          700
        ],
        "desc": "Card MEDIUM: dark rounded panel."
      },
      {
        "type": "text",
        "bbox": [
          260,
          530,
          320,
          680
        ],
        "text": "MEDIUM",
        "desc": "Card title."
      },
      {
        "type": "text",
        "bbox": [
          340,
          530,
          500,
          680
        ],
        "text": "~2.4s at 1K\\n$0.015/image",
        "desc": "Speed and price lines."
      },
      {
        "type": "obj",
        "bbox": [
          240,
          720,
          820,
          920
        ],
        "desc": "Card HIGH: highlighted purple glow border and gradient fill."
      },
      {
        "type": "text",
        "bbox": [
          260,
          750,
          320,
          900
        ],
        "text": "HIGH",
        "desc": "Card title in purple."
      },
      {
        "type": "text",
        "bbox": [
          340,
          750,
          500,
          900
        ],
        "text": "~3.6s at 1K\\n$0.03/image",
        "desc": "Speed and price lines."
      },
      {
        "type": "obj",
        "bbox": [
          860,
          80,
          960,
          920
        ],
        "desc": "Footer bar spanning the bottom."
      },
      {
        "type": "text",
        "bbox": [
          890,
          100,
          940,
          450
        ],
        "text": "Speed \u00b7 Price \u00b7 Quality",
        "desc": "Footer left label."
      },
      {
        "type": "text",
        "bbox": [
          890,
          550,
          940,
          900
        ],
        "text": "trade-off increases \u2192",
        "desc": "Footer right label."
      }
    ]
  }
}

Magazine openers, posters, and covers where the masthead has to stay sharp. Prefer medium or high at 2K when typography is critical.

Magazine opener

media_magazine_opener Medium — **Medium**

Show prompt

{
  "aspect_ratio": "3:4",
  "high_level_description": "A premium vertical magazine feature opener for a culture weekly. Clean white page, bold black masthead at the top, one large photoreal portrait on the right, and a pull-quote block on the left. Prioritize large, crisp, highly legible typography. Avoid tiny body copy and decorative microtext. CRITICAL TYPOGRAPHY: Every visible letter must be perfectly formed and correctly spelled. No overlapping glyphs, no duplicated syllables, no missing/extra characters, no garbled noise, no random digits inside words, and no speed/timing/performance badges or stopwatch callouts. Prefer fewer, larger text elements over dense copy.",
  "compositional_deconstruction": {
    "background": "Matte off-white paper fill with a subtle warm grain. Soft shadow under the portrait. Editorial print look.",
    "global_text_rules": "All visible text must be sharp, correctly spelled, and easy to read at magazine size. Use fewer words and larger font sizes. Do not generate paragraphs or fake barcodes. CRITICAL TYPOGRAPHY: Every visible letter must be perfectly formed and correctly spelled. No overlapping glyphs, no duplicated syllables, no missing/extra characters, no garbled noise, no random digits inside words, and no speed/timing/performance badges or stopwatch callouts. Prefer fewer, larger text elements over dense copy.",
    "elements": [
      {
        "type": "text",
        "bbox": [
          60,
          40,
          120,
          940
        ],
        "text": "ATLAS WEEKLY",
        "desc": "Small uppercase black masthead across the top edge. Tight letter spacing, refined sans-serif."
      },
      {
        "type": "text",
        "bbox": [
          140,
          60,
          220,
          520
        ],
        "text": "ISSUE 48",
        "desc": "Tiny secondary label under the masthead on the left. Light gray."
      },
      {
        "type": "obj",
        "bbox": [
          180,
          420,
          920,
          960
        ],
        "desc": "Large editorial portrait of a fictional woman in her 30s, three-quarter crop, charcoal turtleneck, soft window light, realistic skin and hands, looking slightly off-camera."
      },
      {
        "type": "text",
        "bbox": [
          280,
          70,
          520,
          400
        ],
        "text": "THE\nLAST\nEDIT",
        "desc": "Huge stacked black display title on the left, three short lines, high contrast."
      },
      {
        "type": "text",
        "bbox": [
          560,
          70,
          700,
          400
        ],
        "text": "\"Creativity is a\ncollaboration now.\"",
        "desc": "Medium italic pull-quote in dark charcoal. Two short lines only."
      },
      {
        "type": "text",
        "bbox": [
          740,
          70,
          820,
          400
        ],
        "text": "By Mira Cole",
        "desc": "Small clean byline under the quote."
      }
    ]
  }
}

Podcast cover

Show prompt

Square podcast cover artwork for a fictional interview podcast titled exactly "OFF THE RECORD", hosted by Nina Ellis.

Create a bold editorial studio portrait of a stylish woman framed from the chest upward, posed in a confident three-quarter angle with her head turned slightly to the side. She has a fashion-forward look, short textured platinum-blonde bob, subtle hoop earrings, and a sleek black dress or top with elegant structure. Her expression is poised, thoughtful, and slightly mysterious.

Use a clean, saturated crimson-red background with a soft studio feel, inspired by modern fashion portrait photography. Keep the composition simple, striking, and centered, with the subject occupying most of the frame. The overall aesthetic should feel polished, premium, and highly thumbnail-friendly.

Place the title "OFF THE RECORD" as a tight, cohesive typographic block, not widely separated. Keep "OFF THE" small but close to "RECORD", positioned directly above or aligned closely with it so the full title reads as one unit. Use elegant high-contrast typography: either small uppercase sans-serif for "OFF THE" and large bold serif or refined display typography for "RECORD", or a similarly premium magazine-style combination. The title should feel integrated with the portrait, with large scale and strong hierarchy.

Add the host line exactly: "With Nina Ellis" in smaller clean typography, placed subtly near the bottom or lower side of the composition. Keep it clearly legible but secondary to the main title.

Style direction: luxury editorial portrait, fashion-magazine aesthetic, bold red monochrome palette, clean negative space, minimal clutter, realistic skin texture, soft controlled lighting, sophisticated typography, premium podcast branding.

Exact visible text only:
"OFF THE RECORD"
"With Nina Ellis"
No additional text, no logos, no watermark.

CRITICAL TYPOGRAPHY: Every visible letter must be perfectly formed and correctly spelled. No overlapping glyphs, no duplicated syllables, no missing/extra characters, no garbled noise, no random digits inside words, and no speed/timing/performance badges or stopwatch callouts. Prefer fewer, larger text elements over dense copy.

Series poster

Show prompt

Premium vertical poster for a fictional psychological mystery series titled exactly "THE ECHO ROOM".

Side-profile silhouette of a woman on a textured light-grey background. Inside her silhouette, a double-exposure: a cool teal institutional hallway with a young man's profile in the upper half, and a red-lit corridor with a distant dark figure in the lower half. Moody thriller atmosphere, cinematic grain, high-end streaming key art.

Exact visible text only:
Main title exactly: "THE ECHO ROOM"
Tagline exactly: "Every memory belongs to someone."
No logos, no additional text.

CRITICAL TYPOGRAPHY: Every visible letter must be perfectly formed and correctly spelled. No overlapping glyphs, no duplicated syllables, no missing/extra characters, no garbled noise, no random digits inside words. Prefer fewer, larger text elements over dense copy.

Integration

Endpoint

Base URL: https://api.pruna.ai/v1/predictions

Rate limit: 500 requests per minute

Full API reference: P-Image-Ideogram

Authentication

-H 'apikey: YOUR_API_KEY'

Asynchronous (default)

curl -X POST 'https://api.pruna.ai/v1/predictions' \
-H 'Content-Type: application/json' \
-H 'apikey: YOUR_API_KEY' \
-H 'Model: p-image-ideogram' \
-d '{
    "input": {
        "prompt": "Vertical beauty campaign visual for a fictional fragrance launch...",
        "thinking": "medium",
        "image_size": "2K",
        "aspect_ratio": "3:4",
        "prompt_upsampling": true
    }
}'

Synchronous (Try-Sync)

curl -X POST 'https://api.pruna.ai/v1/predictions' \
-H 'Content-Type: application/json' \
-H 'apikey: YOUR_API_KEY' \
-H 'Model: p-image-ideogram' \
-H 'Try-Sync: true' \
-d '{
    "input": {
        "prompt": "Editorial fashion photograph of a fictional model in Copenhagen rain...",
        "thinking": "low"
    }
}'

Configuration

Parameters below match the live P-Image-Ideogram API reference and follow the same request shape as p-image.

Required parameters

Parameter	Type	Description
prompt	string	Short text prompt for image generation. Supports natural language and JSON structured prompts.

Optional parameters

Parameter	Type	Default	Description
thinking	string	`medium`	Quality level: `very_low`, `low`, `medium`, `high`. Higher thinking → higher quality; pair `high` with detailed prompts and 2K+ for text.
prompt_upsampling	boolean	`true`	Enhance the prompt before generation. Set `false` to pass the raw prompt directly to the model.
aspect_ratio	string	`1:1`	`1:1`, `16:9`, `9:16`, `4:3`, `3:4`, `3:2`, `2:3`, or `custom`
image_size	string	`1K`	Output resolution budget: `1K` or `2K` (ignored when `aspect_ratio=custom`). Use 2K with `high` for best text rendering.
width	integer	`1024`	Custom width (0–2560, multiple of 16). Only when `aspect_ratio=custom`.
height	integer	`1024`	Custom height (0–2560, multiple of 16). Only when `aspect_ratio=custom`.
seed	integer	random	Random seed for reproducible generation
output_format	string	`jpg`	Output image format
output_quality	integer	`80`	Output quality 0–100 (ignored for PNG)

Example: text rendering at High + 2K

{
    "input": {
        "prompt": "Vertical beauty campaign... Exact visible text only: Product name exactly \"ROUGE ATELIER\"...",
        "thinking": "high",
        "image_size": "2K",
        "aspect_ratio": "3:4",
        "prompt_upsampling": true,
        "seed": 42
    }
}

Example: fast draft at Very Low

{
    "input": {
        "prompt": "Editorial fashion photograph of a fictional model walking in Copenhagen rain, no logos, no text.",
        "thinking": "very_low",
        "prompt_upsampling": false
    }
}

Example: raw JSON prompt without upsampling

{
    "input": {
        "prompt": "{\"aspect_ratio\": \"3:4\", \"high_level_description\": \"Luxury fashion lookbook cover...\", \"compositional_deconstruction\": {...}}",
        "thinking": "medium",
        "prompt_upsampling": false
    }
}