World's first open-source, agentic video production system. 12 production pipelines, 100+ tools, 700+ agent skill and production-knowledge files. Turn your AI coding assistant into a full video production studio.
Python agentagentic-aiaiclaudecopilotcursorelevenlabsffmpegfluximage-generationopen-sourceopenaipythonremotionstable-diffusiontext-to-speechtext-to-videovideo-generationvideo-production
Replace retired OpenAI image models with gpt-image-2 (DALL-E shut down 2026-05-12)
Remotely Humanfabc0e70886e
diff --git a/.env.example b/.env.exampleindex f5233118..ffad5a8c 100644--- a/.env.example+++ b/.env.example@@ -16,7 +16,7 @@ GOOGLE_CLOUD_LOCATION= # Vertex AI region, default us-central1# --- Voice ---ELEVENLABS_API_KEY= # TTS narration, music generation, sound effects-OPENAI_API_KEY= # OpenAI TTS fallback and DALL-E image generation+OPENAI_API_KEY= # OpenAI TTS fallback and GPT Image 2 image generationXAI_API_KEY= # Grok image generation/editing and Grok video generationDOUBAO_SPEECH_API_KEY= # Volcengine Doubao Speech TTS (new console API Key)DOUBAO_SPEECH_VOICE_TYPE= # Default Doubao speaker/voice type, e.g. zh_female_vv_uranus_bigttsdiff --git a/AGENT_GUIDE.md b/AGENT_GUIDE.mdindex 085a52d3..4d374d28 100644--- a/AGENT_GUIDE.md+++ b/AGENT_GUIDE.md@@ -489,7 +489,7 @@ Three selector tools abstract multi-provider capabilities. **Selectors auto-disc| Selector | Routes to | How it discovers ||----------|-----------|-----------------|| `tts_selector` | All tools with `capability="tts"` (ElevenLabs, Google TTS, OpenAI, Piper) | `registry.get_by_capability("tts")` |-| `image_selector` | All tools with `capability="image_generation"` (FLUX, Google Imagen, DALL-E, Recraft, etc.) | `registry.get_by_capability("image_generation")` |+| `image_selector` | All tools with `capability="image_generation"` (FLUX, Google Imagen, GPT Image, Recraft, etc.) | `registry.get_by_capability("image_generation")` || `video_selector` | All tools with `capability="video_generation"` | `registry.get_by_capability("video_generation")` |Selectors route based on: user preference > availability > discovery order. They adapt input schemas between providers transparently.diff --git a/README.md b/README.mdindex 168f4ca2..c53432ca 100644--- a/README.md+++ b/README.md@@ -183,7 +183,7 @@ SUNO_API_KEY=your-key # Full songs, instrumentals, any genre# Voice & images:ELEVENLABS_API_KEY=your-key # Premium TTS, AI music, sound effects-OPENAI_API_KEY=your-key # OpenAI TTS, DALL-E 3 images+OPENAI_API_KEY=your-key # OpenAI TTS, GPT Image 2 imagesXAI_API_KEY=your-key # xAI Grok image edits/generation + Grok video generationGOOGLE_API_KEY=your-key # Google Imagen images, Google TTS (700+ voices)@@ -447,7 +447,7 @@ Each tool declares which Layer 3 skills it relies on. The agent reads Layer 1 to| **FLUX** | Cloud API | State-of-the-art quality || **Google Imagen** | Cloud API | Imagen 4 — high-quality, multiple aspect ratios || **Grok Imagine Image** | Cloud API | Strong image edits, style transfer, and multi-image compositing |-| **DALL-E 3** | Cloud API | OpenAI's image model |+| **GPT Image 2** | Cloud API | OpenAI's image model || **Recraft** | Cloud API | Design-focused generation || **Local Diffusion** | Local GPU | Stable Diffusion, free || **Pexels** | Stock | Free stock images |diff --git a/README_zh-CN.md b/README_zh-CN.mdindex e962b0c3..ca663047 100644--- a/README_zh-CN.md+++ b/README_zh-CN.md@@ -177,7 +177,7 @@ SUNO_API_KEY=your-key # 完整的歌曲、伴奏,涵盖任何流派# 语音与图像:ELEVENLABS_API_KEY=your-key # 顶级 TTS、AI 音乐、音效-OPENAI_API_KEY=your-key # OpenAI TTS、DALL-E 3 图像+OPENAI_API_KEY=your-key # OpenAI TTS、GPT Image 2 图像XAI_API_KEY=your-key # xAI Grok 图像编辑/生成 + Grok 视频生成GOOGLE_API_KEY=your-key # Google Imagen 图像、Google TTS(700+ 种声音)@@ -441,7 +441,7 @@ OpenMontage/| **FLUX** | 云端 API | 业界顶尖质量 || **Google Imagen** | 云端 API | Imagen 4 — 高质量、多种长宽比 || **Grok Imagine Image** | 云端 API | 强大的图像编辑、风格转换和多图合成 |-| **DALL-E 3** | 云端 API | OpenAI 的图像模型 |+| **GPT Image 2** | 云端 API | OpenAI 的图像模型 || **Recraft** | 云端 API | 专注于设计的生成 || **Local Diffusion** | 本地 GPU | Stable Diffusion,免费 || **Pexels** | 素材库 | 免费的库存图片 |diff --git a/docs/ARCHITECTURE.md b/docs/ARCHITECTURE.mdindex e7b26d94..a983956c 100644--- a/docs/ARCHITECTURE.md+++ b/docs/ARCHITECTURE.md@@ -48,7 +48,7 @@ OpenMontage/│ ├── audio/ # TTS (ElevenLabs, OpenAI, Piper), music gen, mixing, enhancement│ ├── avatar/ # Talking head animation, lip sync│ ├── enhancement/ # Upscale, bg removal, face enhance/restore, color grading-│ ├── graphics/ # Image gen (FLUX, DALL-E, Recraft, local diffusion), stock, diagrams, code snippets, math animation+│ ├── graphics/ # Image gen (FLUX, GPT Image, Recraft, local diffusion), stock, diagrams, code snippets, math animation│ ├── publishers/ # (Reserved)│ ├── subtitle/ # SRT/VTT generation from timestamps│ └── video/ # 13 video gen providers, composition, stitching, trimming@@ -383,7 +383,7 @@ All config is validated via Pydantic models in `lib/config_model.py`.| Variable | Used By | Purpose ||----------|---------|---------|| `ELEVENLABS_API_KEY` | elevenlabs_tts, music_gen | TTS, music, sound effects |-| `OPENAI_API_KEY` | openai_tts, openai_image | TTS fallback, DALL-E 3 |+| `OPENAI_API_KEY` | openai_tts, openai_image | TTS fallback, GPT Image 2 || `XAI_API_KEY` | grok_image, grok_video | Grok image editing/generation, Grok video generation || `FAL_KEY` | flux_image, kling_video, veo_video, minimax_video, recraft_image | fal.ai hosted models (FLUX, Veo, Kling, MiniMax, Recraft) || `HEYGEN_API_KEY` | heygen_video | Multi-provider video generation |diff --git a/docs/PROVIDERS.md b/docs/PROVIDERS.mdindex f0460a69..5aa809d4 100644--- a/docs/PROVIDERS.md+++ b/docs/PROVIDERS.md@@ -15,7 +15,7 @@ Everything you need to know about every provider in OpenMontage — setup instru| 3 | **$0** | ElevenLabs | Premium TTS + music + SFX (10K chars/month free) || 4 | **$0** | Piper (local install) | Fully offline TTS — no API key, no cost, no network || 5 | **~$0.03/image** | fal.ai | FLUX images + Kling/Veo/MiniMax video + Recraft — broad single-key image + video coverage |-| 6 | **~$0.04/image** | OpenAI | DALL-E 3 images + OpenAI TTS |+| 6 | **~$0.05/image** | OpenAI | GPT Image 2 images + OpenAI TTS || 7 | **~$0.04/image** | Google Imagen | Imagen 4 images (shares the Google API key) || 8 | **$12/month** | Runway | Gen-4 video — highest quality AI video || 9 | **pay-as-you-go** | HeyGen | Avatar videos, multi-model video gateway |@@ -37,7 +37,7 @@ GOOGLE_API_KEY= # Google TTS + Google Imagen# VOICE + MUSICELEVENLABS_API_KEY= # TTS, music, sound effects (10K chars/month free)-OPENAI_API_KEY= # OpenAI TTS + DALL-E 3 images+OPENAI_API_KEY= # OpenAI TTS + GPT Image 2 imagesXAI_API_KEY= # xAI Grok image generation/editing + Grok video generationDOUBAO_SPEECH_API_KEY= # Volcengine Doubao Speech TTS (strong Mandarin narration)DOUBAO_SPEECH_VOICE_TYPE= # Default Doubao speaker/voice type@@ -276,7 +276,7 @@ Google TTS offers 700+ voices across 50+ languages. Voice names follow the patte### OpenAI — TTS + Image Generation-> **Solid all-rounder.** DALL-E 3 handles complex multi-element compositions well. TTS is fast and affordable.+> **Solid all-rounder.** GPT Image 2 handles complex multi-element compositions and in-image text well. TTS is fast and affordable.**Tools unlocked:** `openai_tts`, `openai_image`**Env var:** `OPENAI_API_KEY`@@ -301,10 +301,14 @@ Google TTS offers 700+ voices across 50+ languages. Voice names follow the patte| Model | Size | Quality | Price per image ||-------|------|---------|----------------|-| DALL-E 3 | 1024x1024 | standard | $0.040 |-| DALL-E 3 | 1024x1024 | hd | $0.080 |-| DALL-E 3 | 1024x1792 | standard | $0.080 |-| DALL-E 3 | 1024x1792 | hd | $0.120 |+| GPT Image 2 | 1024x1024 | low | $0.006 |+| GPT Image 2 | 1024x1024 | medium | $0.053 |+| GPT Image 2 | 1024x1024 | high | $0.211 |+| GPT Image 2 | 1024x1536 / 1536x1024 | low | $0.005 |+| GPT Image 2 | 1024x1536 / 1536x1024 | medium | $0.041 |+| GPT Image 2 | 1024x1536 / 1536x1024 | high | $0.165 |++> **Note:** DALL-E 2/3 were shut down by OpenAI on 2026-05-12, and the `gpt-image-1` family (`gpt-image-1-mini`, `gpt-image-1.5`) retires 2026-12-01 — `gpt-image-2` is OpenAI's recommended replacement ([deprecations](https://developers.openai.com/api/docs/deprecations)).**Free tier:** None. Requires prepaid billing. Previously offered $5 in free credits for new accounts (discontinued for most signups).@@ -672,7 +676,7 @@ First run downloads the model (~4GB). Subsequent runs use the cached model.**VRAM requirement:** 4GB+ (8GB recommended for 1024x1024 images)-**Supports:** Negative prompts, seeds, custom sizes. Quality is lower than FLUX or DALL-E 3 but completely free and offline.+**Supports:** Negative prompts, seeds, custom sizes. Quality is lower than FLUX or GPT Image 2 but completely free and offline.---@@ -743,7 +747,7 @@ How many providers cover each capability:| Capability | Cloud Providers | Local Providers | Free Options ||-----------|----------------|-----------------|--------------|-| **Image Generation** | FLUX, Grok, Google Imagen, DALL-E 3, Recraft | Local Diffusion | Pexels, Pixabay (stock) |+| **Image Generation** | FLUX, Grok, Google Imagen, GPT Image 2, Recraft | Local Diffusion | Pexels, Pixabay (stock) || **Video Generation** | Grok, Kling, Runway, Veo, Higgsfield, MiniMax, HeyGen | WAN, Hunyuan, CogVideo, LTX | Pexels, Pixabay (stock) || **Text-to-Speech** | ElevenLabs, Google TTS, OpenAI | Piper | Piper, Google free tier, ElevenLabs free tier || **Music Generation** | ElevenLabs, Suno | — | ElevenLabs free tier |diff --git a/skills/creative/data-visualization.md b/skills/creative/data-visualization.mdindex 73d4bdd0..7648d3d9 100644--- a/skills/creative/data-visualization.md+++ b/skills/creative/data-visualization.md@@ -12,7 +12,7 @@ clear, accurate, and effective in video.| Tool | Role ||------|------|| `diagram_gen` | Generate charts via Mermaid or D3 |-| `image_selector` | Generate stylized chart illustrations (FLUX/DALL-E) |+| `image_selector` | Generate stylized chart illustrations (FLUX/GPT Image) || Remotion | Animated chart components (bar grow, line draw, pie fill) || Manim | Mathematical plots, coordinate systems, function graphs |diff --git a/skills/creative/image-gen-usage.md b/skills/creative/image-gen-usage.mdindex 018de149..3b263d40 100644--- a/skills/creative/image-gen-usage.md+++ b/skills/creative/image-gen-usage.md@@ -1,6 +1,6 @@# Image Generation Usage for OpenMontage-> Sources: OpenAI DALL-E 3 documentation, FLUX/BFL API documentation, existing Layer 3 skills+> Sources: OpenAI GPT Image documentation, FLUX/BFL API documentation, existing Layer 3 skills> at `.agents/skills/flux-best-practices/` and `.agents/skills/bfl-api/`## Quick Reference Card@@ -150,7 +150,7 @@ optimized for image/video generation providers.## Common Pitfalls1. **Text in images** — AI image generators are unreliable with text. Never include text in prompts; add text as overlays in the compose stage-2. **Hands and fingers** — DALL-E 3 and FLUX still struggle. Avoid prompts requiring detailed hand poses+2. **Hands and fingers** — AI image models still struggle. Avoid prompts requiring detailed hand poses3. **Inconsistent characters** — Without reference images, the same character will look different each time. Always use the hero reference strategy4. **Over-prompting** — Long, complex prompts produce unpredictable results. Keep to 2-3 sentences5. **Over-unifying prompts** — Forcing the exact same style phrase into every prompt makes scenes look samey. Keep the visual system consistent, but let each scene express its own subject, shot, and emotional beat.diff --git a/skills/creative/image-provider-usage.md b/skills/creative/image-provider-usage.mdindex 179e64c0..47f842a3 100644--- a/skills/creative/image-provider-usage.md+++ b/skills/creative/image-provider-usage.md@@ -11,7 +11,7 @@|------|----------|------|-------|----------|| `flux_image` | FLUX 2 Pro via fal.ai | ~$0.03-0.05 | ~5-10s | Photorealism, general purpose, workhorse || `grok_image` | Grok Imagine Image (xAI) | $0.02/output + $0.002/input edit image | ~5-15s | Image edits, style transfer, multi-image compositing |-| `openai_image` | GPT Image 1 (OpenAI) | ~$0.01-0.17 | ~5-15s | Complex instructions, text in images, multi-element |+| `openai_image` | GPT Image 2 (OpenAI) | ~$0.01-0.21 | ~5-15s | Complex instructions, text in images, multi-element || `recraft_image` | Recraft V4 via fal.ai | ~$0.04-0.25 | ~5-10s | Logos, SVG vectors, brand assets, text rendering (see caveat below) || `local_diffusion` | Stable Diffusion (local) | Free | ~30s+ | Offline, privacy, free || `image_gen` | Multi (legacy, deprecated) | Varies | Varies | **Deprecated** — use `image_selector` or per-provider tools |@@ -39,7 +39,7 @@| **Style transfer / repaint of an existing image** | `grok_image` | Native edit flow, strong promptable transforms | `openai_image` || **Multi-image merge / composite** | `grok_image` | Can combine multiple source images into one scene | `openai_image` || **Logo or brand asset** | `recraft_image` | SVG support, text accuracy | `openai_image` |-| **Image with text/labels** | `openai_image` | Best text rendering (GPT Image 1) | `recraft_image` |+| **Image with text/labels** | `openai_image` | Best text rendering (GPT Image 2) | `recraft_image` || **Complex multi-element composition** | `openai_image` | Best instruction following | `flux_image` || **Hero image (key visual)** | `flux_image` | Highest visual quality | `openai_image` || **Thumbnail** | `flux_image` or `recraft_image` | Needs to be eye-catching | — |@@ -59,7 +59,7 @@PRODUCTION PATH: Premium├── Hero images: flux_image ($0.05/img)├── Supporting visuals: flux_image ($0.03/img)-├── Text overlays: openai_image ($0.04/img)+├── Text overlays: openai_image ($0.05/img medium)├── B-roll stills: pexels_image ($0.00)└── Total for 10 images: ~$0.35diff --git a/skills/pipelines/animation/asset-director.md b/skills/pipelines/animation/asset-director.mdindex 26830c00..d8bbf55a 100644--- a/skills/pipelines/animation/asset-director.md+++ b/skills/pipelines/animation/asset-director.md@@ -159,7 +159,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/animation/proposal-director.md b/skills/pipelines/animation/proposal-director.mdindex 5dc7bdeb..41d21579 100644--- a/skills/pipelines/animation/proposal-director.md+++ b/skills/pipelines/animation/proposal-director.md@@ -278,7 +278,7 @@ For each concept, specify:For each concept, specify:- **Animation approach**: `image_animation` / `clip_video` / `manim` / `remotion_dataviz` / `diagram_stills` / `mixed`- **Why this approach**: grounded in technique research AND tool availability from Step 3-- **Image/video generation provider**: which specific provider from the preflight scan (e.g., "FLUX via fal.ai", "gpt-image-1 via OpenAI", "Stable Diffusion local")+- **Image/video generation provider**: which specific provider from the preflight scan (e.g., "FLUX via fal.ai", "gpt-image-2 via OpenAI", "Stable Diffusion local")- **Reuse strategy**: What's the visual system? (recurring motifs, layout grid, color scheme, transition family)- **Complexity estimate**: How many unique scene types vs. reusable templates?- **Visual identity**: palette, typography, texture, motion energy, and why they fit this subject, audience, and platform@@ -461,7 +461,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/avatar-spokesperson/asset-director.md b/skills/pipelines/avatar-spokesperson/asset-director.mdindex 9b2ca32d..1600e3b2 100644--- a/skills/pipelines/avatar-spokesperson/asset-director.md+++ b/skills/pipelines/avatar-spokesperson/asset-director.md@@ -124,7 +124,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/cinematic/asset-director.md b/skills/pipelines/cinematic/asset-director.mdindex 73f45a26..036eae8e 100644--- a/skills/pipelines/cinematic/asset-director.md+++ b/skills/pipelines/cinematic/asset-director.md@@ -157,7 +157,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/cinematic/proposal-director.md b/skills/pipelines/cinematic/proposal-director.mdindex 9508e164..d40a2a70 100644--- a/skills/pipelines/cinematic/proposal-director.md+++ b/skills/pipelines/cinematic/proposal-director.md@@ -285,7 +285,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/clip-factory/asset-director.md b/skills/pipelines/clip-factory/asset-director.mdindex 027cc605..6655f211 100644--- a/skills/pipelines/clip-factory/asset-director.md+++ b/skills/pipelines/clip-factory/asset-director.md@@ -101,7 +101,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/explainer/asset-director.md b/skills/pipelines/explainer/asset-director.mdindex b98c358d..2215e8a5 100644--- a/skills/pipelines/explainer/asset-director.md+++ b/skills/pipelines/explainer/asset-director.md@@ -273,7 +273,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/explainer/proposal-director.md b/skills/pipelines/explainer/proposal-director.mdindex 526cc61c..086970c8 100644--- a/skills/pipelines/explainer/proposal-director.md+++ b/skills/pipelines/explainer/proposal-director.md@@ -535,7 +535,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/explainer/scene-director.md b/skills/pipelines/explainer/scene-director.mdindex 0d37827a..2141ce5b 100644--- a/skills/pipelines/explainer/scene-director.md+++ b/skills/pipelines/explainer/scene-director.md@@ -79,7 +79,7 @@ Transform each script section into 1-3 visual scenes. Each scene is a distinct v| `text_card` | Statements, closing messages, key terms | Remotion TextCard (centered, spring animation) | 3-5s || `animation` | Concepts needing motion (data flow, math) | Remotion, Manim | 4-10s || `diagram` | Processes, architecture, relationships | `diagram_gen` (Mermaid), `image_selector` | 4-8s |-| `generated` | Illustrations, metaphors, real-world imagery | `image_selector` (FLUX/DALL-E) | 3-6s |+| `generated` | Illustrations, metaphors, real-world imagery | `image_selector` (FLUX/GPT Image) | 3-6s || `talking_head` | AI avatar speaking (if HeyGen available) | HeyGen tools | 5-15s || `broll` | Context, real-world examples | Stock or generated footage | 3-6s || `screen_recording` | Code demos, UI walkthroughs | Recorded or simulated | 5-15s |@@ -206,7 +206,7 @@ The style playbook constrains your visual choices:**Feasibility check:**- [ ] Every `required_asset` with `source: "generate"` is achievable with available tools- [ ] Diagram descriptions are specific enough for Mermaid syntax generation-- [ ] Image descriptions are specific enough for FLUX/DALL-E prompt engineering+- [ ] Image descriptions are specific enough for FLUX/GPT Image prompt engineering- [ ] No scene requires tools that aren't in the tool registry### Step 7: Self-Evaluatediff --git a/skills/pipelines/hybrid/asset-director.md b/skills/pipelines/hybrid/asset-director.mdindex 89b7e0e5..8213e389 100644--- a/skills/pipelines/hybrid/asset-director.md+++ b/skills/pipelines/hybrid/asset-director.md@@ -93,7 +93,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/localization-dub/asset-director.md b/skills/pipelines/localization-dub/asset-director.mdindex 637f3d44..18d7e562 100644--- a/skills/pipelines/localization-dub/asset-director.md+++ b/skills/pipelines/localization-dub/asset-director.md@@ -85,7 +85,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/podcast-repurpose/asset-director.md b/skills/pipelines/podcast-repurpose/asset-director.mdindex 2902ad2c..d78df75b 100644--- a/skills/pipelines/podcast-repurpose/asset-director.md+++ b/skills/pipelines/podcast-repurpose/asset-director.md@@ -98,7 +98,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/screen-demo/asset-director.md b/skills/pipelines/screen-demo/asset-director.mdindex 4f3df502..770683c9 100644--- a/skills/pipelines/screen-demo/asset-director.md+++ b/skills/pipelines/screen-demo/asset-director.md@@ -159,7 +159,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/skills/pipelines/talking-head/asset-director.md b/skills/pipelines/talking-head/asset-director.mdindex c0082607..80c58bd8 100644--- a/skills/pipelines/talking-head/asset-director.md+++ b/skills/pipelines/talking-head/asset-director.md@@ -201,7 +201,7 @@ If you encounter a generation technique, provider behavior, or prompting patternThis is especially important for:- **Video generation prompting** — models respond to specific vocabularies that change with each version-- **Image model parameters** — optimal settings for FLUX, DALL-E, Imagen differ and evolve+- **Image model parameters** — optimal settings for FLUX, GPT Image, Imagen differ and evolve- **Audio provider quirks** — voice cloning, music generation, and TTS each have model-specific best practices- **Remotion component patterns** — new composition techniques emerge as the framework evolvesdiff --git a/tests/qa/QA_PLAN.md b/tests/qa/QA_PLAN.mdindex d2656307..961fef86 100644--- a/tests/qa/QA_PLAN.md+++ b/tests/qa/QA_PLAN.md@@ -9,7 +9,7 @@ Run every tool with real API keys, inspect outputs (see images, listen to audio,| Script | Tools Tested | API Keys Used | Est. Cost ||--------|-------------|---------------|-----------|| `test_01_tts.py` | `elevenlabs_tts` (ElevenLabs) | ELEVENLABS_API_KEY | ~$0.02 |-| `test_02_image_gen.py` | `image_gen` (DALL-E 3 + FLUX) | OPENAI_API_KEY, FAL_AI_API_KEY | ~$0.15 |+| `test_02_image_gen.py` | `image_gen` (GPT Image 2 + FLUX) | OPENAI_API_KEY, FAL_AI_API_KEY | ~$0.15 || `test_03_music.py` | `music_gen` (ElevenLabs) | ELEVENLABS_API_KEY | ~$0.10 || `test_04_audio_mix.py` | `audio_mixer` | None (ffmpeg only) | $0 || `test_05_video_compose.py` | `video_compose` | None (ffmpeg only) | $0 |@@ -30,7 +30,7 @@ For each output:| Area | Risk | How to Validate ||------|------|-----------------|| TTS voice selection | Default voice may not match playbook mood | Test with multiple voice IDs, compare against playbook `voice_style` |-| Image gen consistency | DALL-E/FLUX outputs vary wildly per prompt | Test with playbook `image_prompt_prefix` prepended |+| Image gen consistency | GPT Image/FLUX outputs vary wildly per prompt | Test with playbook `image_prompt_prefix` prepended || Music duration alignment | Music may not match narration duration | Compare `music.duration` vs `tts.duration`, check padding/looping || Audio ducking timing | Ducking may cut music too aggressively | Inspect waveform: music should duck ~6dB under speech, recover smoothly || Video stitch transitions | Crossfade may flicker with mismatched codecs | Test with both matching and mismatched clips, check `auto_normalize` |diff --git a/tools/graphics/image_gen.py b/tools/graphics/image_gen.pyindex 933e0922..11e14884 100644--- a/tools/graphics/image_gen.py+++ b/tools/graphics/image_gen.py@@ -7,7 +7,7 @@pexels_image, pixabay_image). This file is kept for backwardscompatibility and will be removed in a future release.-Supports cloud API providers (FLUX via fal.ai/Replicate, OpenAI DALL-E)+Supports cloud API providers (FLUX via fal.ai/Replicate, OpenAI GPT Image)and local Stable Diffusion via diffusers. Reports unavailable withinstall instructions when no provider is configured."""@@ -43,12 +43,12 @@ class ImageGen(BaseTool):stability = ToolStability.EXPERIMENTALexecution_mode = ExecutionMode.SYNCdeterminism = Determinism.SEEDED- runtime = ToolRuntime.HYBRID # API (DALL-E/FLUX) or local (diffusers)+ runtime = ToolRuntime.HYBRID # API (GPT Image/FLUX) or local (diffusers)dependencies = [] # checked dynamically based on providerinstall_instructions = ("Set one of these environment variables:\n"- " OPENAI_API_KEY — for DALL-E 3\n"+ " OPENAI_API_KEY — for GPT Image 2\n"" FAL_KEY — for FLUX via fal.ai\n""Or install diffusers for local generation:\n"" pip install diffusers transformers accelerate torch"@@ -121,7 +121,7 @@ class ImageGen(BaseTool):def estimate_cost(self, inputs: dict[str, Any]) -> float:provider = inputs.get("provider") or self._detect_provider()if provider == "openai":- return 0.04 # DALL-E 3 standard+ return 0.053 # gpt-image-2 medium at 1024x1024 (call uses auto quality)if provider == "flux":return 0.03return 0.0 # local@@ -159,14 +159,14 @@ class ImageGen(BaseTool):client = OpenAI()prompt = inputs["prompt"]size = f"{inputs.get('width', 1024)}x{inputs.get('height', 1024)}"- model = inputs.get("model", "dall-e-3")+ model = inputs.get("model", "gpt-image-2")+ # GPT image models don't accept response_format; they always return b64response = client.images.generate(model=model,prompt=prompt,size=size,n=1,- response_format="b64_json",)image_data = base64.b64decode(response.data[0].b64_json)diff --git a/tools/graphics/openai_image.py b/tools/graphics/openai_image.pyindex 5c5cd8fa..cabf2c1e 100644--- a/tools/graphics/openai_image.py+++ b/tools/graphics/openai_image.py@@ -1,4 +1,4 @@-"""OpenAI GPT Image generation (gpt-image-2 / DALL-E 3)."""+"""OpenAI GPT Image generation (gpt-image-2)."""from __future__ import annotations@@ -60,20 +60,17 @@ class OpenAIImage(BaseTool):"prompt": {"type": "string"},"model": {"type": "string",- "enum": ["gpt-image-2", "dall-e-3"],+ "enum": ["gpt-image-2"],"default": "gpt-image-2",},"size": {"type": "string",- "enum": [- "1024x1024", "1536x1024", "1024x1536", "auto",- "1024x1792", "1792x1024", # dall-e-3 only- ],+ "enum": ["1024x1024", "1536x1024", "1024x1536", "auto"],"default": "1024x1024",},"quality": {"type": "string",- "enum": ["low", "medium", "high", "auto", "standard", "hd"],+ "enum": ["low", "medium", "high", "auto"],"default": "high",},"output_format": {@@ -100,15 +97,12 @@ class OpenAIImage(BaseTool):return ToolStatus.UNAVAILABLEdef estimate_cost(self, inputs: dict[str, Any]) -> float:- model = inputs.get("model", "gpt-image-2")+ # gpt-image-2 per-image pricing at 1024x1024 (non-square sizes run+ # slightly cheaper): https://developers.openai.com/api/docs/guides/image-generationquality = inputs.get("quality", "high")n = inputs.get("n", 1)- if "gpt-image" in model:- cost_map = {"low": 0.011, "medium": 0.042, "high": 0.167, "auto": 0.042}- return cost_map.get(quality, 0.042) * n- # dall-e-3 fallback pricing- quality_map = {"standard": 0.04, "hd": 0.08}- return quality_map.get(quality, 0.04) * n+ cost_map = {"low": 0.006, "medium": 0.053, "high": 0.211, "auto": 0.053}+ return cost_map.get(quality, 0.053) * ndef execute(self, inputs: dict[str, Any]) -> ToolResult:if not os.environ.get("OPENAI_API_KEY"):@@ -127,30 +121,16 @@ class OpenAIImage(BaseTool):n = inputs.get("n", 1)try:- if "gpt-image" in model:- quality = inputs.get("quality", "high")- output_format = inputs.get("output_format", "png")- response = client.images.generate(- model=model,- prompt=prompt,- size=size,- quality=quality,- output_format=output_format,- n=n,- )- else:- # dall-e-3 path- quality = inputs.get("quality", "standard")- if quality in ("low", "medium", "high", "auto"):- quality = "standard" # map to dall-e-3 quality options- response = client.images.generate(- model=model,- prompt=prompt,- size=size,- quality=quality,- n=1, # dall-e-3 only supports n=1- response_format="b64_json",- )+ quality = inputs.get("quality", "high")+ output_format = inputs.get("output_format", "png")+ response = client.images.generate(+ model=model,+ prompt=prompt,+ size=size,+ quality=quality,+ output_format=output_format,+ n=n,+ )image_data = base64.b64decode(response.data[0].b64_json)ext = inputs.get("output_format", "png")