Content Creator Studio

Multi-agent AI content creation platform built with LangGraph, FastAPI, and a progressively refactored DDD architecture.

Overview • Architecture • Quick Start • API • Layout

English | 中文

---

Overview

Content Creator Studio is an AI-assisted content workflow for solo creators and small teams. The backend uses LangGraph to orchestrate agent execution, FastAPI to expose HTTP APIs, and a progressively introduced DDD structure to separate domain logic, application services, adapters, and infrastructure.

The repository now reflects a refactor in progress: the new architecture is live in the codebase, while selected legacy paths are preserved for compatibility and migration.

Current Architecture

Active agents

The current graph routes requests to 3 core agents:

Agent	Best for	Notes
SimpleAgent	quick chat and lightweight tasks	fastest path
ReActAgent	search, tool use, factual tasks	integrates tool calling and knowledge search
ReflectionAgent	high-quality long-form output	supports refinement and deeper reasoning

Key changes from the older design

• PlanSolveAgent is no longer part of the active main flow; its capability is being consolidated into ReflectionAgent.
• RAGAgent has been downgraded into a reusable tool, KnowledgeSearchTool.
• Routing is moving from hard-coded logic to strategy-based configuration via app/config/routing_config.yaml.
• Memory recall and persistence are centralized through application services such as ContextBuilder and MemoryAppService.
• Legacy compatibility is still available under app/agents/_legacy/ and app/adapters/.

Main capabilities

• LangGraph-driven execution with routing, quality gate, refinement, and memory save steps
• Long-term memory for chat, content, and knowledge flows
• Knowledge retrieval backed by in-memory vectors or Milvus
• Streaming APIs, video generation workflow, and observability endpoints
• Configurable external tool integration through the tool registry and MCP bridge

Quick Start

Prerequisites

• Python 3.11+
• An API key compatible with the configured LLM provider
• Optional: Redis and Milvus for extended features

Run locally

git clone https://github.com/kingdoja/content-creator-studio.git
cd content-creator-studio/iccp-langchain

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

cp .env.example .env
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

On Windows PowerShell:

python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -r requirements.txt
Copy-Item .env.example .env
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

After startup:

• App root: http://localhost:8000/
• OpenAPI docs: http://localhost:8000/docs
• Health check: http://localhost:8000/health

API Surface

Important API groups currently exposed from app.main:

• /api/v1/content for content creation, cover generation, evaluation, and video-related flows
• /api/v1/chat for chat sessions and messaging
• /api/v1/knowledge for knowledge base operations
• /api/v1/memory for recall and preference APIs
• /api/v1/observability and /api/v1/metrics for monitoring
• /api/v1/auth for authentication and WeChat-related auth routes

Repository Layout

onePersonCompany/
├── iccp-langchain/           # Main FastAPI + LangGraph application
├── docs/                     # Local project notes (ignored from Git)
├── ideas/                    # Local ideation notes (ignored from Git)
├── miniprogram/              # Local mini-program work area (ignored from Git)
├── web/                      # Local web experiments (ignored from Git)
└── _local_only/              # Local-only files (ignored from Git)

Inside iccp-langchain/:

app/
├── agents/                   # graph, router_v2, analyzers, core agents, legacy agents
├── adapters/                 # compatibility adapters
├── api/v1/                   # HTTP routes
├── config/                   # settings and routing config
├── domain/                   # domain models and interfaces
├── memory/                   # memory infra
├── observability/            # tracing and monitoring helpers
├── rag/                      # embeddings and vector index
├── services/                 # application services
└── tools/                    # web search, fact check, knowledge search, MCP bridge
tests/                        # automated tests
scripts/                      # migration and verification utilities

Notes For Contributors

• The repo intentionally ignores local AI-tooling directories and generated cover images.
• Runtime-generated assets should stay out of version control unless there is a strong product reason to ship them.
• The most detailed backend documentation is in iccp-langchain/README.md.
• Release notes for the architecture refactor live in iccp-langchain/CHANGELOG.md.

License

MIT