Skip to content

AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He

Add MiniMax image backend support

hugohe39dd16514e22b
diff --git a/.env.example b/.env.example
index 1d580be0..3e3d6aa9 100644
--- a/.env.example
+++ b/.env.example
@@ -38,6 +38,14 @@
# OPENAI_MODEL=gpt-image-1
# OPENAI_BASE_URL=http://127.0.0.1:3000/v1
+# ─────────────────────────────────────────────────────────────
+# Example: MiniMax / MiniMax 示例
+# ─────────────────────────────────────────────────────────────
+# IMAGE_BACKEND=minimax
+# MINIMAX_API_KEY=your-minimax-key
+# MINIMAX_MODEL=image-01
+# MINIMAX_BASE_URL=https://api.minimaxi.com/v1/image_generation
+
# ─────────────────────────────────────────────────────────────
# Example: Qwen / 通义生图示例
# ─────────────────────────────────────────────────────────────
@@ -78,6 +86,10 @@
# IDEOGRAM_MODEL=ideogram-v3
# IDEOGRAM_BASE_URL=https://api.ideogram.ai
#
+# MINIMAX_API_KEY=your-minimax-key
+# MINIMAX_MODEL=image-01
+# MINIMAX_BASE_URL=https://api.minimaxi.com/v1/image_generation
+#
# SILICONFLOW_API_KEY=your-siliconflow-key
# SILICONFLOW_MODEL=Qwen/Qwen-Image
# SILICONFLOW_BASE_URL=https://api.siliconflow.cn/v1/images/generations
diff --git a/skills/ppt-master/scripts/docs/image.md b/skills/ppt-master/scripts/docs/image.md
index 459bacf3..b7810eff 100644
--- a/skills/ppt-master/scripts/docs/image.md
+++ b/skills/ppt-master/scripts/docs/image.md
@@ -17,12 +17,13 @@ python3 scripts/image_gen.py --list-backends
Support tiers:
- Core: `gemini`, `openai`, `qwen`, `zhipu`, `volcengine`
- Extended: `stability`, `bfl`, `ideogram`
-- Experimental: `siliconflow`, `fal`, `replicate`
+- Experimental: `minimax`, `siliconflow`, `fal`, `replicate`
Backend selection:
```bash
python3 scripts/image_gen.py "A cat" --backend openai
+python3 scripts/image_gen.py "A cinematic portrait" --backend minimax
python3 scripts/image_gen.py "A product launch hero image" --backend qwen
python3 scripts/image_gen.py "科技感背景图" --backend zhipu
python3 scripts/image_gen.py "A product KV in cinematic style" --backend volcengine
@@ -54,7 +55,7 @@ export GEMINI_MODEL=gemini-3.1-flash-image-preview
Current process environment wins over `.env`.
-Use provider-specific keys only, such as `GEMINI_API_KEY`, `OPENAI_API_KEY`, `QWEN_API_KEY`, `ZHIPU_API_KEY`, `VOLCENGINE_API_KEY`, `FAL_KEY`, or `REPLICATE_API_TOKEN`.
+Use provider-specific keys only, such as `GEMINI_API_KEY`, `OPENAI_API_KEY`, `MINIMAX_API_KEY`, `QWEN_API_KEY`, `ZHIPU_API_KEY`, `VOLCENGINE_API_KEY`, `FAL_KEY`, or `REPLICATE_API_TOKEN`.
`IMAGE_API_KEY`, `IMAGE_MODEL`, and `IMAGE_BASE_URL` are intentionally unsupported.
diff --git a/skills/ppt-master/scripts/image_backends/backend_minimax.py b/skills/ppt-master/scripts/image_backends/backend_minimax.py
new file mode 100644
index 00000000..1be33ba2
--- /dev/null
+++ b/skills/ppt-master/scripts/image_backends/backend_minimax.py
@@ -0,0 +1,197 @@
+#!/usr/bin/env python3
+"""
+MiniMax image generation backend.
+
+Configuration keys:
+ MINIMAX_API_KEY (required)
+ MINIMAX_BASE_URL (optional)
+ MINIMAX_MODEL (optional)
+"""
+
+import base64
+import os
+import time
+
+import requests
+
+from image_backends.backend_common import (
+ MAX_RETRIES,
+ detect_image_extension,
+ http_error,
+ is_rate_limit_error,
+ normalize_image_size,
+ require_api_key,
+ resolve_output_path,
+ retry_delay,
+ save_image_bytes,
+)
+
+
+DEFAULT_ENDPOINT = "https://api.minimaxi.com/v1/image_generation"
+DEFAULT_MODEL = "image-01"
+
+ASPECT_RATIO_SIZE_MAP = {
+ "512px": {
+ "1:1": (512, 512),
+ "16:9": (640, 360),
+ "4:3": (576, 432),
+ "3:2": (624, 416),
+ "2:3": (416, 624),
+ "3:4": (432, 576),
+ "9:16": (360, 640),
+ "21:9": (672, 288),
+ },
+ "1K": {
+ "1:1": (1024, 1024),
+ "16:9": (1280, 720),
+ "4:3": (1152, 864),
+ "3:2": (1248, 832),
+ "2:3": (832, 1248),
+ "3:4": (864, 1152),
+ "9:16": (720, 1280),
+ "21:9": (1344, 576),
+ },
+ "2K": {
+ "1:1": (2048, 2048),
+ "16:9": (2048, 1152),
+ "4:3": (2048, 1536),
+ "3:2": (2048, 1368),
+ "2:3": (1368, 2048),
+ "3:4": (1536, 2048),
+ "9:16": (1152, 2048),
+ "21:9": (2048, 880),
+ },
+ "4K": {
+ "1:1": (2048, 2048),
+ "16:9": (2048, 1152),
+ "4:3": (2048, 1536),
+ "3:2": (2048, 1368),
+ "2:3": (1368, 2048),
+ "3:4": (1536, 2048),
+ "9:16": (1152, 2048),
+ "21:9": (2048, 880),
+ },
+}
+
+
+def _resolve_url(base_url: str) -> str:
+ """Resolve the MiniMax image generation endpoint."""
+ base = base_url.rstrip("/")
+ if base.endswith("/image_generation"):
+ return base
+ return base + "/v1/image_generation"
+
+
+def _resolve_dimensions(aspect_ratio: str, image_size: str) -> tuple[int, int]:
+ """Resolve width and height from the unified aspect_ratio/image_size pair."""
+ normalized = normalize_image_size(image_size)
+ dimensions = (ASPECT_RATIO_SIZE_MAP.get(normalized) or {}).get(aspect_ratio)
+ if not dimensions:
+ supported = sorted(ASPECT_RATIO_SIZE_MAP["1K"])
+ raise ValueError(
+ f"Unsupported aspect ratio '{aspect_ratio}' for MiniMax backend. "
+ f"Supported: {supported}"
+ )
+ return dimensions
+
+
+def _extract_image_bytes(payload: dict) -> bytes | None:
+ """Extract image bytes from a MiniMax response payload."""
+ data = payload.get("data") or {}
+ image_base64 = data.get("image_base64") or []
+ if image_base64:
+ return base64.b64decode(image_base64[0])
+ return None
+
+
+def _generate_image(api_key: str, prompt: str, negative_prompt: str = None,
+ aspect_ratio: str = "1:1", image_size: str = "1K",
+ output_dir: str = None, filename: str = None,
+ model: str = DEFAULT_MODEL, base_url: str = DEFAULT_ENDPOINT) -> str:
+ """Generate one image with the MiniMax backend."""
+ width, height = _resolve_dimensions(aspect_ratio, image_size)
+ url = _resolve_url(base_url)
+ final_prompt = prompt
+ if negative_prompt:
+ final_prompt += f"\n\nAvoid the following: {negative_prompt}"
+
+ headers = {
+ "Authorization": f"Bearer {api_key}",
+ "Content-Type": "application/json",
+ }
+ payload = {
+ "model": model,
+ "prompt": final_prompt,
+ "width": width,
+ "height": height,
+ "response_format": "base64",
+ "n": 1,
+ }
+
+ print("[MiniMax Image]")
+ print(f" Model: {model}")
+ print(f" Prompt: {final_prompt[:120]}{'...' if len(final_prompt) > 120 else ''}")
+ print(f" Aspect Ratio: {aspect_ratio}")
+ print(f" Resolution: {width}x{height} (from image_size={image_size})")
+ print()
+ print(" [..] Generating...", end="", flush=True)
+ start = time.time()
+ response = requests.post(url, headers=headers, json=payload, timeout=300)
+ elapsed = time.time() - start
+ print(f"\n [DONE] Response received ({elapsed:.1f}s)")
+
+ if response.status_code != 200:
+ raise http_error(response, "MiniMax image generation")
+
+ data = response.json()
+ base_resp = data.get("base_resp") or {}
+ if base_resp.get("status_code") not in (None, 0, "0"):
+ raise RuntimeError(f"MiniMax image generation failed: {data}")
+
+ image_bytes = _extract_image_bytes(data)
+ if not image_bytes:
+ raise RuntimeError(f"MiniMax response missing image data: {data}")
+
+ ext = detect_image_extension(image_bytes) or ".jpeg"
+ path = resolve_output_path(prompt, output_dir, filename, ext)
+ return save_image_bytes(image_bytes, path)
+
+
+def generate(prompt: str, negative_prompt: str = None,
+ aspect_ratio: str = "1:1", image_size: str = "1K",
+ output_dir: str = None, filename: str = None,
+ model: str = None, max_retries: int = MAX_RETRIES) -> str:
+ """Generate an image with retries using the MiniMax backend."""
+ api_key = require_api_key(
+ "MINIMAX_API_KEY",
+ message="No API key found. Set MINIMAX_API_KEY in the current environment or the project-root .env.",
+ )
+ base_url = os.environ.get("MINIMAX_BASE_URL") or DEFAULT_ENDPOINT
+ resolved_model = model or os.environ.get("MINIMAX_MODEL") or DEFAULT_MODEL
+ normalized_size = normalize_image_size(image_size)
+
+ last_error = None
+ for attempt in range(max_retries + 1):
+ try:
+ return _generate_image(
+ api_key=api_key,
+ prompt=prompt,
+ negative_prompt=negative_prompt,
+ aspect_ratio=aspect_ratio,
+ image_size=normalized_size,
+ output_dir=output_dir,
+ filename=filename,
+ model=resolved_model,
+ base_url=base_url,
+ )
+ except Exception as exc:
+ last_error = exc
+ if attempt >= max_retries:
+ break
+ limited = is_rate_limit_error(exc)
+ delay = retry_delay(attempt, rate_limited=limited)
+ label = "Rate limit hit" if limited else f"Error: {exc}"
+ print(f"\n [WARN] {label}. Retrying in {delay}s...")
+ time.sleep(delay)
+
+ raise RuntimeError(f"Failed after {max_retries + 1} attempts. Last error: {last_error}")
diff --git a/skills/ppt-master/scripts/image_gen.py b/skills/ppt-master/scripts/image_gen.py
index 0f7226ea..b964b042 100644
--- a/skills/ppt-master/scripts/image_gen.py
+++ b/skills/ppt-master/scripts/image_gen.py
@@ -7,6 +7,7 @@ Dispatches to the appropriate backend based on explicit provider configuration.
Backend selection (`IMAGE_BACKEND` in `.env` or the current process environment):
IMAGE_BACKEND=gemini -> Gemini backend (google-genai SDK)
IMAGE_BACKEND=openai -> OpenAI-compatible backend (openai SDK)
+ IMAGE_BACKEND=minimax -> MiniMax image backend
IMAGE_BACKEND=stability -> Stability AI backend
IMAGE_BACKEND=bfl -> Black Forest Labs FLUX backend
IMAGE_BACKEND=ideogram -> Ideogram backend
@@ -45,6 +46,7 @@ IMAGE_ENV_PREFIXES = (
"IMAGE_",
"GEMINI_",
"OPENAI_",
+ "MINIMAX_",
"STABILITY_",
"BFL_",
"IDEOGRAM_",
@@ -91,6 +93,14 @@ BACKEND_REGISTRY = {
"key_hint": "OPENAI_API_KEY",
"aliases": ["openai-compatible", "openai_compatible"],
},
+ "minimax": {
+ "module": "backend_minimax",
+ "tier": "experimental",
+ "label": "MiniMax Image",
+ "default_model": "image-01",
+ "key_hint": "MINIMAX_API_KEY",
+ "aliases": ["minimaxi"],
+ },
"qwen": {
"module": "backend_qwen",
"tier": "core",