topic-to-outline ai text generation with structured json output
Converts a user-provided text topic into a structured content outline by routing requests through pluggable AI text generation clients (Google GenAI, OpenAI-compatible APIs). The system uses a provider configuration abstraction layer to support multiple LLM backends, with prompt engineering that enforces JSON schema compliance for downstream image generation. Implements retry mechanisms and error handling to ensure reliable outline generation even with transient API failures.
Unique: Uses a provider-agnostic configuration system (provider_config.yaml) that abstracts text generation clients, allowing runtime swapping between Google GenAI, OpenAI, and OpenAI-compatible APIs without code changes. Implements structured prompt engineering with JSON schema validation to ensure outline output is deterministic and directly consumable by the image generation pipeline.
vs alternatives: More flexible than single-provider solutions (e.g., Copilot, ChatGPT API) because it decouples LLM selection from application code, enabling cost optimization and provider failover without redeployment.
multi-provider image generation with consistent visual styling
Generates 6-9 styled images from outline content by orchestrating multiple image generation backends (Google GenAI, Banana.dev Nano Pro, OpenAI-compatible APIs) through an abstraction layer. Each image is generated with embedded Chinese text, consistent visual design across the series, and optional reference image conditioning. The system applies image compression and optimization post-generation to reduce file sizes while maintaining quality for social media distribution.
Unique: Implements a pluggable image generator architecture with three distinct backends (GoogleGenAIGenerator, ImageAPIGenerator for Banana.dev, OpenAICompatibleGenerator) that share a common interface, enabling provider-agnostic image generation. Includes post-generation image compression and optimization specifically tuned for Xiaohongshu's platform constraints (aspect ratios, file size limits).
vs alternatives: Supports specialized image generation providers (Banana.dev Nano Pro) optimized for fast, cost-effective generation, whereas generic tools like Midjourney or DALL-E lack platform-specific optimization and require manual post-processing for social media formats.
chinese text rendering and embedding in generated images
Embeds Chinese text directly into generated images during the image generation phase, using LLM-based text generation (outline content) and provider-specific text rendering capabilities. The system generates Chinese text via the outline generation phase, passes it to image generation prompts with explicit text embedding instructions, and validates that generated images contain readable Chinese text. Handles character encoding (UTF-8), font selection, and text layout to ensure accurate Chinese text rendering without post-generation OCR or manual text addition.
Unique: Integrates Chinese text generation (outline phase) with image generation (image phase) to embed text directly in generated images via LLM prompts, avoiding post-processing steps. Relies on image generation model's instruction-following to accurately render Chinese text.
vs alternatives: More integrated than tools requiring separate text overlay or OCR steps; faster than manual design because text is embedded during generation rather than added post-hoc, but less reliable than explicit font rendering because it depends on LLM instruction-following.
api routes and rest endpoints for generation workflow
Exposes Flask REST API endpoints for the two-phase generation workflow: POST /api/generate/outline (topic → outline), POST /api/generate/images (outline → images), and GET /api/generate/status (progress polling). Each endpoint accepts JSON request bodies with generation parameters (topic, reference images, provider config), validates inputs, and returns JSON responses with generated content or error details. Implements request validation, error handling, and optional authentication/rate limiting for production deployments.
Unique: Implements Flask REST API endpoints for the two-phase generation workflow (outline → images), with SSE streaming for progress updates and JSON request/response format for easy integration.
vs alternatives: More flexible than web-only interfaces because it exposes programmatic API access, enabling third-party integrations and automation; simpler than GraphQL for this use case because REST is sufficient for the linear generation workflow.
reference image multimodal conditioning for content generation
Accepts optional user-uploaded reference images and incorporates them into both outline generation and image generation pipelines via multimodal LLM APIs. The system encodes reference images as base64 or file uploads, passes them to text and image generation models that support vision capabilities, and uses them to influence content style, tone, and visual direction without explicit fine-tuning. Handles image validation, format conversion, and size constraints before submission to downstream providers.
Unique: Integrates reference image handling directly into the content generation pipeline (both outline and image phases) via multimodal LLM APIs, rather than as a post-processing step. Abstracts image encoding and validation to support multiple provider APIs (Google GenAI, OpenAI) with different image submission formats.
vs alternatives: More integrated than tools requiring separate style transfer or LoRA fine-tuning steps; reference images influence generation in real-time without additional training, making it faster for one-off or low-volume content creation.
server-sent events (sse) streaming for real-time generation progress
Streams generation progress updates to the frontend in real-time using HTTP Server-Sent Events (SSE), allowing users to monitor outline generation and image generation phases without polling. The backend emits progress events at key checkpoints (outline started, outline completed, image 1 generated, image 2 generated, etc.), and the frontend Vue.js application listens to these events and updates the UI reactively. Enables long-running operations (30+ seconds) to feel responsive and transparent to users.
Unique: Implements SSE streaming at the Flask application level, emitting progress events from both outline generation and image generation phases, with frontend Vue.js components listening to EventSource and updating UI reactively via Pinia state management.
vs alternatives: More efficient than polling-based progress tracking (which adds unnecessary API calls) and simpler than WebSocket for one-directional server-to-client updates; native browser support via EventSource API requires no additional libraries.
provider configuration abstraction with runtime provider swapping
Implements a configuration-driven provider selection system where text and image generation providers are specified in YAML/JSON configuration files (provider_config.yaml) rather than hardcoded in application logic. At runtime, the system instantiates the appropriate text/image generator client based on configuration, enabling users to swap providers (Google GenAI → OpenAI → Ollama) without code changes or redeployment. Configuration includes API endpoints, model names, authentication credentials, and provider-specific parameters (temperature, max_tokens, image resolution).
Unique: Uses a provider-agnostic factory pattern where TextGenerationClient and ImageGeneratorClient are abstract base classes, with concrete implementations (GoogleGenAITextClient, OpenAITextClient, OllamaTextClient, etc.) instantiated based on configuration at application startup. Configuration is externalized to YAML, decoupling provider selection from application code.
vs alternatives: More flexible than single-provider tools (ChatGPT, Midjourney) because provider selection is configuration-driven rather than hardcoded, enabling cost optimization and provider failover without code changes or redeployment.
image compression and optimization for social media distribution
Automatically compresses and optimizes generated images post-generation to meet Xiaohongshu platform constraints (file size, aspect ratio, resolution). The system applies lossy/lossless compression algorithms, generates thumbnail variants, and validates output dimensions and file sizes before returning to user. Compression parameters are tunable via configuration to balance quality vs. file size based on platform requirements.
Unique: Implements post-generation image optimization specifically tuned for Xiaohongshu's platform constraints (aspect ratios, file size limits), with configurable compression parameters and automatic thumbnail generation for gallery display.
vs alternatives: More integrated than external image optimization tools (ImageMagick, TinyPNG) because compression is built into the generation pipeline and tuned for Xiaohongshu's specific requirements, eliminating manual post-processing steps.
+4 more capabilities