awesome-ai-agents

Production-ready AI agent examples you can clone and run in 5 minutes.
Every example in both Python and TypeScript. No fluff, just code that works.

Getting Started • Examples • Contributing

---

Getting Started

git clone https://github.com/vakra-dev/awesome-ai-agents.git
cd awesome-ai-agents

Pick an example, follow its README. That's it. No monorepo setup, no shared dependencies, no build step.

How This Repo Is Organized

awesome-ai-agents/
  agents/
    starter/       -- Your first agents. Simple, focused, one concept each.
    advanced/      -- Production patterns. Multi-step, parallel, complex.
  multi-agent/     -- Teams of agents working together.
  rag/             -- Retrieval-Augmented Generation patterns.
  mcp/             -- Model Context Protocol servers and clients.
  memory/          -- Persistent agent memory patterns.
  voice/           -- Voice and realtime agents.
  frameworks/      -- Framework-specific quickstarts.
  patterns/        -- Reusable agent patterns (streaming, retry, HITL).

Every example has the same structure:

example-name/
  README.md           -- What it does, how to set up, how it works
  python/
    main.py           -- Run with: python main.py
    requirements.txt  -- pip install -r requirements.txt
    .env.example      -- Copy to .env and add your API keys
  typescript/
    index.ts          -- Run with: npx tsx index.ts
    package.json      -- npm install
    .env.example      -- Copy to .env and add your API keys

Examples

AI Agents -- Starter

Simple, focused agents. One concept each. Great for learning the basics.

#	Example	What It Does	Models
1	Research Agent	Web research with structured markdown reports	OpenAI
2	Code Review Agent	Multi-input code review with severity ratings	Anthropic
3	Data Analyst Agent	CSV analysis with code generation and execution	OpenAI
4	Web Scraping Agent	Structured data extraction from any URL	Anthropic
5	Customer Support Agent	RAG-based support with escalation logic	OpenAI
6	Paper Summarizer	Research paper summarization via Reader	OpenAI + Reader
7	Content Repurposer	Turn blog posts into tweets, LinkedIn posts, emails	Anthropic + Reader
8	Newsletter Curator	Read multiple blogs, curate a newsletter digest	OpenAI + Reader
9	Fact-Checker Agent	Verify claims against web sources with verdicts	Anthropic + Reader
10	Git Commit Agent	Generate conventional commit messages from diffs	Anthropic
11	Regex Generator	Natural language to tested regex patterns	OpenAI
12	Cron Translator	Bidirectional cron expression translation	OpenAI
13	Dockerfile Generator	Read a project, generate optimized Dockerfile	Anthropic
14	SQL Agent	Natural language queries against SQLite	OpenAI
15	JSON Transformer	Transform JSON with natural language instructions	OpenAI
16	Email Drafter	Draft professional emails with tone control	Anthropic
17	Dependency Audit	Check dependencies for known vulnerabilities	OpenAI
18	Secret Scanner	Scan a codebase for leaked credentials	Anthropic
19	Standup Summarizer	Generate standup updates from git log	OpenAI

AI Agents -- Advanced

Production-grade agents. Multi-step, parallel execution, complex orchestration.

#	Example	What It Does	Models
20	Deep Research Agent	Multi-step research with parallel search and cost tracking	Anthropic
21	Coding Agent	Plan-code-test-fix loop with filesystem tools	Anthropic
22	Computer Use Agent	Browser control via screenshots and Playwright	Anthropic
23	Competitor Monitor	Track competitor site changes over time via Reader	Anthropic + Reader
24	Lead Enrichment	Extract structured company intel from websites via Reader	Anthropic + Reader
25	SEO Audit Agent	Crawl and analyze a site for SEO issues via Reader	OpenAI + Reader
26	PR Review Agent	Fetch PR diffs from GitHub, review like a senior engineer	Anthropic
27	Log Analyzer	Find anomalies and root causes in log files	Anthropic
28	Incident Responder	Triage alerts, produce incident response plans	OpenAI
29	API Test Generator	Generate test suites from OpenAPI specs	Anthropic
30	Prompt Optimizer	Iteratively improve prompts with test cases	OpenAI
31	Eval Runner	Compare multiple models on the same task	Multi-model
32	Fine-tune Data Generator	Generate synthetic training data for fine-tuning	Anthropic

Multi-Agent Teams

Multiple agents collaborating in pipelines with handoffs and feedback loops.

#	Example	What It Does	Models
33	Software Dev Team	PM, Architect, Developer, Reviewer pipeline	OpenAI
34	Content Pipeline	Researcher, Writer, Editor with revision loops	Anthropic

RAG (Retrieval-Augmented Generation)

Ground LLM responses in real data using vector search and retrieval.

#	Example	What It Does	Models
35	PDF Chatbot	Upload a PDF, ask questions conversationally	OpenAI
36	Codebase RAG	Index and query a codebase with code-aware chunking	Anthropic
37	Agentic RAG	Smart routing: retrieve vs answer vs clarify	OpenAI
38	Documentation Q&A	Crawl any docs site via Reader, build RAG index, answer questions	OpenAI + Reader

MCP (Model Context Protocol)

MCP servers and clients for tool-based integrations.

#	Example	What It Does	Models
39	MCP Postgres Server	Expose PostgreSQL as MCP tools	--
40	MCP GitHub Server	GitHub operations as MCP tools	--
41	MCP Client Agent	Multi-server MCP client with tool routing	OpenAI

Memory

Persistent agent memory across sessions and conversations.

#	Example	What It Does	Models
42	Conversation Memory	Remember conversations across sessions with SQLite	OpenAI
43	Entity Memory	Track entities and relationships over time	Anthropic

Voice

Real-time voice and audio agents.

#	Example	What It Does	Models
44	Realtime Voice Agent	Real-time conversation via OpenAI Realtime API	OpenAI Realtime

Frameworks

Framework-specific quickstarts and examples.

#	Example	What It Does	Models
45	OpenAI Agents SDK	5 sub-examples: basic, tools, handoffs, guardrails, full	OpenAI

Patterns

Reusable agent patterns you can drop into any project.

#	Example	What It Does	Models
46	Human-in-the-Loop	Pause for approval before risky actions	OpenAI
47	Streaming	Stream text and tool calls in real-time	Anthropic
48	Retry and Fallback	Exponential backoff with model fallback	Multi-model

API Keys You'll Need

Most examples need one or more API keys. Here's the full list:

Provider	Key	Used By	Get One
OpenAI	OPENAI_API_KEY	25+ examples	platform.openai.com
Anthropic	ANTHROPIC_API_KEY	20+ examples	console.anthropic.com
Reader	READER_API_KEY	8 examples (web reading)	reader.dev
Tavily	TAVILY_API_KEY	5 examples (web search)	tavily.com
GitHub	GITHUB_TOKEN	3 examples (GitHub API)	github.com/settings/tokens

Contributing

We welcome contributions! Please read our Contributing Guide before submitting a PR.

Every new example must:

• Include both Python and TypeScript implementations
• Follow the example template
• Run in under 5 minutes
• Include a README following our template

License

MIT -- see LICENSE for details.

---

Star this repo to stay updated with new examples!