Repolode
Submit
NirDiamant

NirDiamant/agents-towards-production

A collection of 28 code-first tutorials for taking GenAI agents from prototype to production, covering memory, RAG, deployment, security and observability.

View on GitHubWebsite
Stars
21,732
Forks
2,880
Open issues
5
Last commit
3 days ago

What is it?

Agents Towards Production is a set of runnable tutorials, each in its own folder with notebooks or code files, for building production-grade GenAI agents. The topics span the components of an agent stack: stateful workflows with LangGraph, vector and dual-memory storage with Redis and Mem0, web search and scraping via Tavily and Bright Data, MCP-based tool integration, and secure tool calling with OAuth2 and human approval.

The deployment side covers Docker containers, FastAPI endpoints with streaming, local models through Ollama, AWS Bedrock AgentCore, and GPU deployment on RunPod. Further tutorials address prompt-injection guardrails, LangSmith tracing, agent evaluation, fine-tuning, multi-agent communication over the A2A protocol, Kotlin agents built with JetBrains' Koog, and a Streamlit chat UI.

Several tutorials are sponsored by the companies whose tools they use, and the repository also links to a paid course by its author.

Key features

  • 01Covers stateful agent workflows with LangGraph and multi-agent communication via A2A.
  • 02Includes memory tutorials using Redis, Mem0 and Cognee for persistent and semantic recall.
  • 03Shows deployment paths from Docker and FastAPI to Ollama, AWS Bedrock AgentCore and RunPod GPUs.
  • 04Provides security tutorials on prompt-injection guardrails and secure OAuth2 tool calling.
  • 05Adds observability and evaluation tutorials using LangSmith and IntellAgent.

Use cases

  • Moving a prototype LLM agent toward a containerized, API-served deployment.
  • Adding persistent memory or RAG ingestion pipelines to an existing agent.
  • Learning how to add guardrails and tracing before exposing an agent to users.
  • Comparing managed platforms and self-hosted options for running agents.

Things to know

  • The license is not clearly identified (listed as NOASSERTION), so reuse terms are unclear.

Topics

#ai-agents#llm#langgraph#rag#agent-memory#mcp#deployment#observability#tutorials#agent

FAQ

What is agents-towards-production?
A collection of 28 code-first tutorials for taking GenAI agents from prototype to production, covering memory, RAG, deployment, security and observability. Agents Towards Production is a set of runnable tutorials, each in its own folder with notebooks or code files, for building production-grade GenAI agents. The topics span the components of an agent stack: stateful workflows with LangGraph, vector and dual-memory storage with Redis and Mem0, web search and scraping via Tavily and Bright Data, MCP-based tool integration, and secure tool calling with OAuth2 and human approval.
Is agents-towards-production free to use? What license does it have?
agents-towards-production is open source under the Other license. Check the license terms before using it in commercial products.
Is agents-towards-production still maintained?
Actively maintained. The last commit to agents-towards-production was on Oct 17, 2026.
What can I use agents-towards-production for?
Moving a prototype LLM agent toward a containerized, API-served deployment. Adding persistent memory or RAG ingestion pipelines to an existing agent. Learning how to add guardrails and tracing before exposing an agent to users. Comparing managed platforms and self-hosted options for running agents.
Where can I get agents-towards-production?
The source code and installation instructions are on GitHub: https://github.com/NirDiamant/agents-towards-production

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects