10 ready-to-use agents · CLI + API · Extensible framework · OpenAI / OpenRouter compatible
Quick Start · Agents · API Server · Build Your Own · Roadmap
AgentHub is an open-source library of AI agents that automate real-world tasks — SEO analysis, market research, content creation, lead generation, and more. Each agent is production-ready, modular, and can be used via CLI, API, or imported as a Python module.
No complex setup. No boilerplate. Just run.
python run_agent.py seo "Analyze the SEO of https://yoursite.com"
python run_agent.py startup "Generate AI startup ideas for healthcare"
python run_agent.py blog "Write a blog post about autonomous agents"- 10 Specialized Agents — Each agent is an expert at a specific task
- One-Line Execution — Run any agent with a single CLI command
- REST API — FastAPI server to integrate agents into any workflow
- Tool System — Web search, scraping, file I/O, API calls built-in
- Memory System — Agents remember context across reasoning steps
- OpenRouter / OpenAI Compatible — Use any LLM provider
- Modular Architecture — Easy to extend, customize, or chain agents
- Pipeline Support — Chain multiple agents for complex workflows
- Auto Retry — Exponential backoff on LLM failures
- Step Callbacks — Monitor agent reasoning in real-time
- Structured Logging — Built-in Python logging for debugging
- Docker Ready — One-command deployment with Docker Compose
git clone https://github.com/yourusername/AgentHub.git
cd AgentHub
pip install -r requirements.txt# Option A: OpenRouter (recommended – access to 100+ models)
export OPENROUTER_API_KEY=your_key_here
# Option B: OpenAI directly
export OPENAI_API_KEY=your_key_herepython run_agent.py seo "Analyze the SEO of https://example.com"That's it. 🎉
| Agent | Key | Description |
|---|---|---|
| 🔍 SEO Analyzer | seo |
Full website SEO audit with actionable recommendations |
| 💡 Startup Idea Generator | startup |
Generates and validates startup ideas with market analysis |
| 🐦 Twitter Content | twitter |
Creates content calendars, tweets, and thread strategies |
| 📊 Market Research | market |
In-depth market analysis with SWOT and competitive landscape |
| 🔧 Code Review | code_review |
Reviews code for bugs, security issues, and best practices |
| ✍️ Blog Writer | blog |
SEO-optimized, publication-ready blog posts |
| 🎯 Lead Generation | leads |
Finds and qualifies B2B leads with outreach suggestions |
| 📈 Trend Finder | trends |
Discovers emerging trends with impact analysis |
| ⚔️ Competitor Analyzer | competitor |
Deep competitive analysis with feature comparison |
| 🚀 Product Launch | launch |
Complete launch plans with go-to-market strategies |
# CLI
python run_agent.py <agent_key> "Your task description"
# Examples
python run_agent.py seo "Audit https://mysite.com for SEO issues"
python run_agent.py startup "Generate fintech startup ideas in Latin America"
python run_agent.py twitter "Create a week of Twitter content for a dev tools brand"
python run_agent.py market "Research the no-code platform market"
python run_agent.py code_review "Review the code in ./src/main.py"
python run_agent.py blog "Write about microservices vs monoliths"
python run_agent.py leads "Find leads for a B2B email marketing tool"
python run_agent.py trends "What are the emerging trends in Web3?"
python run_agent.py competitor "Compare Figma vs Sketch vs Adobe XD"
python run_agent.py launch "Plan the launch of a new AI coding assistant"AgentHub includes a FastAPI server for REST API access:
# Start the server
uvicorn server:app --reload
# Run an agent via API
curl -X POST http://localhost:8000/agent/seo \
-H "Content-Type: application/json" \
-d '{"task": "Analyze SEO of https://example.com"}'
# List available agents
curl http://localhost:8000/agents| Method | Endpoint | Description |
|---|---|---|
GET |
/health |
Health check |
GET |
/agents |
List all available agents |
POST |
/agent/{name} |
Run an agent with a task |
Chain multiple agents for complex workflows:
from agenthub.agents.trend_finder_agent import TrendFinderAgent
from agenthub.agents.startup_idea_agent import StartupIdeaAgent
from agenthub.agents.competitor_analyzer_agent import CompetitorAnalyzerAgent
# Step 1: Find trends
trends = TrendFinderAgent().run("Emerging AI trends in 2026")
# Step 2: Generate ideas based on trends
ideas = StartupIdeaAgent().run(f"Ideas based on: {trends.output}")
# Step 3: Analyze competition
analysis = CompetitorAnalyzerAgent().run(f"Competitors for: {ideas.output[:1000]}")See the examples/ folder for complete pipeline scripts.
Built-in tools available to all agents:
| Tool | Description |
|---|---|
web_search |
Search the web via DuckDuckGo (no API key needed) |
scrape_url |
Fetch and extract text from any webpage |
read_file |
Read local text files |
write_file |
Write content to local files |
call_api |
Make HTTP requests to external APIs |
Creating a new agent takes ~20 lines:
from agenthub.core import Agent
from agenthub.tools import web_search_tool, write_file_tool
class MyCustomAgent(Agent):
name = "My Custom Agent"
description = "Does something amazing."
system_prompt = """You are an expert at ..."""
max_steps = 8
def __init__(self, **kwargs):
super().__init__(
tools=[web_search_tool, write_file_tool],
**kwargs,
)
# Run it
agent = MyCustomAgent()
result = agent.run("Do the thing")
print(result.output)from agenthub.core.tools import Tool
def my_tool_fn(query: str) -> str:
return f"Result for: {query}"
my_tool = Tool(
name="my_tool",
description="Description for the LLM.",
parameters={
"type": "object",
"properties": {
"query": {"type": "string", "description": "The input query."},
},
"required": ["query"],
},
fn=my_tool_fn,
)agenthub/
├── agents/
│ ├── seo_agent/ # SEO Website Analyzer
│ ├── startup_idea_agent/ # Startup Idea Generator
│ ├── twitter_agent/ # Twitter Content Creator
│ ├── market_research_agent/ # Market Research
│ ├── code_review_agent/ # Code Reviewer
│ ├── blog_writer_agent/ # Blog Writer
│ ├── lead_generation_agent/ # Lead Generator
│ ├── trend_finder_agent/ # Trend Finder
│ ├── competitor_analyzer_agent/ # Competitor Analyzer
│ └── product_launch_agent/ # Product Launcher
├── core/
│ ├── agent.py # Base Agent class
│ ├── tools.py # Tool system
│ └── memory.py # Memory system
├── tools/
│ ├── web_search.py # Web search tool
│ ├── scraper.py # Web scraper tool
│ ├── file_tools.py # File read/write tools
│ └── api_caller.py # HTTP API tool
examples/
├── startup_research.py # Multi-agent pipeline
├── seo_audit.py # SEO audit example
└── content_pipeline.py # Content marketing pipeline
run_agent.py # CLI runner
server.py # FastAPI server
tests/ # Test suite (pytest)
Dockerfile # Container build
docker-compose.yml # One-command deploy
# Build and run with Docker Compose
docker compose up --build
# Or with plain Docker
docker build -t agenthub .
docker run -p 8000:8000 --env-file .env agenthubThe API will be available at http://localhost:8000.
# Install dev dependencies
pip install -e ".[dev]"
# Run tests
pytest tests/ -v- Core agent framework with tool calling
- 10 specialized agents
- CLI runner
- FastAPI REST API
- Memory system
- Docker support
- Automatic retry with backoff
- Step callbacks for monitoring
- Comprehensive test suite
- Streaming responses
- Agent-to-agent communication
- Persistent vector memory (ChromaDB / Qdrant)
- Web UI dashboard
- Authentication & rate limiting for API
- More agents: Email Agent, Data Analyst, Resume Builder, Legal Reviewer
- Plugin system for community agents
- LangSmith / Langfuse tracing integration
We welcome contributions! Here's how:
- Fork the repository
- Create a feature branch:
git checkout -b feat/my-agent - Add your agent in
agenthub/agents/your_agent/ - Test it works:
python run_agent.py your_agent "test task" - Submit a pull request
See CONTRIBUTING.md for detailed guidelines.
- New agents (Email, Data Analysis, Resume, Legal, etc.)
- New tools (PDF reader, database connector, etc.)
- UI/Dashboard
- Better memory systems
- Documentation improvements
MIT License — see LICENSE for details.
⭐ Star this repo if you find it useful!
Built with ❤️ by the open-source community.