MarkTechPost AI-Agents-Projects-Tutorials: A Running Notebook Collection for Agentic AI
Multi-agent systems, memory, planning, reasoning loops
At a glance
- What is it?
- AI-Agents-Projects-Tutorials is a public GitHub repository from MarkTechPost that pairs each Jupyter notebook with a corresponding tutorial article, covering multi-agent systems, memory, planning, reasoning loops and tool use across dozens of frameworks and models. It is an aggregator of applied AI coding examples, not a framework itself.
- Who is it for?
- AI-Agents-Projects-Tutorials is most useful for developers who learn by reading and running concrete code alongside explanatory articles. Each notebook links to its companion piece on marktechpost.com, which explains the design decisions.
- Can I use it commercially?
- Not without permission. GitHub finds no licence file in the repository, and without a licence all rights are reserved by default: you may read the code but not reuse it. Check the README, or ask the authors, before using it.
- Is it still maintained?
- Yes. The repository last received commits 3 days ago.
- What is it written in?
- Mainly Jupyter Notebook, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 28, 2026, and from our analysis. They are not legal advice.
Editorial analysis
What this repository is and who it is for
AI-Agents-Projects-Tutorials is a public collection of Jupyter notebooks and Python scripts maintained by MarkTechPost, an AI media and research publication. The README describes it as covering multi-agent systems, memory, planning and reasoning loops. Each entry in the repository links to a notebook or script alongside a corresponding tutorial article on marktechpost.com.
The audience is developers and researchers who want to see working code for a specific AI agent technique, model or framework. It is not a beginner-to-advanced course with a defined syllabus; instead it is an ongoing record of implementations that MarkTechPost has built and written about. If you want to understand how a specific tool works in practice, such as Kimi CLI for agentic coding, policy-governed multi-agent workflows with Omnigent, or how to use Claude with Python and MCP connectors, this repository gives you a runnable starting point alongside the article explanation.
The repository does not provide a unified framework or reusable library. The notebooks are independent examples. Each one installs its own dependencies in the notebook environment and runs in isolation.
Repository organization and how to navigate it
The top-level directory contains folders organized by theme. The README shows entries from several directories including:
- Agentic AI Codes/: multi-agent workflows, agent runtimes, memory systems - AI Agents Codes/: specific agent implementations and tool-calling examples - Data Analysis/: data science agents and analytical notebooks - Agentic Workflows/: web crawling, pipeline examples, data processing - LLM Projects/: LLM-native programming patterns - MCP Codes/: Model Context Protocol implementations
Alongside these directories, several notebooks live at the repository root for broadly applicable topics: AutoGen with SemanticKernel and Gemini Flash, LangGraph multi-agent research teams, CrewAI workflows and a JSON prompting guide.
The top-level entries also include topical subdirectories for Computer Vision, NLP, RAG, Quantum Computing, Federated Learning and more, indicating that the collection has expanded well beyond agentic AI into general ML.
To use the repository, clone it:
git clone https://github.com/MARKTECHPOST-AI-MEDIA-INC/AI-Agents-Projects-Tutorials.gitThen navigate to the relevant folder and open the notebook of interest in Jupyter or upload it to Google Colab. Each notebook is designed to run in Colab with minimal local setup.
What the notebooks cover: recent examples
The README lists dozens of entries. Recent additions from 2026 include:
A guide to TypeSafe AI Jev: a pattern for typed decisions, calibrated confidence and speculative fan-out with a System One model. The notebook is typesafe_jev_system_one_typed_decisions_tutorial_Marktechpost.ipynb.
Building non-interactive agentic coding workflows with Moonshot AI's Kimi CLI using JSONL streaming, testing and session memory. The notebook is Kimi_CLI_Advanced_Agentic_Coding_Marktechpost.ipynb.
A policy-governed multi-agent financial research workflow with Omnigent: omnigent_multi_agent_fx_research_Marktechpost.ipynb.
Designing skill-driven financial analysis agents with Claude, Python and MCP connectors using a Python script: claude_financial_services_agentic_workflows_marktechpost.py.
Building a T4-friendly autonomous data science agent with DeepAnalyze-8B and sandboxed code execution: deepanalyze_8b_autonomous_data_science_agent_Marktechpost.ipynb.
A Nanobot-style AI agent in Google Colab with tool calling, session memory, skills and MCP servers: nanobot_style_personal_ai_agent_from_scratch_Marktechpost.ipynb.
Each entry follows the same pattern: a link to the notebook in the repository and a link to the tutorial article on marktechpost.com. The notebooks tend to be specific to one tool or framework combination rather than general-purpose templates.
Limitations of this collection as a learning resource
The notebooks are snapshots. An API or library that changes after the notebook was written may cause the code to fail without any indication in the notebook itself. There is no CI to verify that notebooks still run. The repository has no tagged releases and no changelog, so there is no way to know when a notebook was last verified.
Dependency management is notebook-local. Each notebook installs its own packages at runtime, which can conflict with local environments. The recommended approach is to run them in Colab where each session starts clean.
The collection is broad but not deep on any single topic. A developer who needs comprehensive coverage of, say, LangGraph memory patterns would find a few relevant notebooks here but would need to consult the LangGraph documentation and upstream examples for completeness.
The repository includes a .DS_Store file at the top level, which is an artifact from macOS directory browsing and suggests the repository was created without a comprehensive .gitignore covering macOS system files.
The README is a flat list of notebook links. There is no searchable index, no difficulty ratings and no guidance on which notebook to start with. Navigation requires reading the list or browsing the directory tree.
How AI-Agents-Projects-Tutorials compares to LangChain's own documentation
LangChain's official documentation and GitHub organization provide reference implementations, quickstart notebooks and how-to guides for LangChain and LangGraph. The official examples are versioned, tested against current library versions and updated when the API changes.
AI-Agents-Projects-Tutorials takes a different approach: it covers many frameworks, not just LangChain, and pairs each notebook with an editorial article that explains the design rationale in addition to the code. It covers Omnigent, Kimi CLI, DeepAnalyze-8B, OpenSpace, Fable 5 and other tools that the official LangChain documentation does not include.
The trade-off is currency versus breadth. Official documentation is more likely to be accurate against the current library version. This collection is more likely to show you a real integration pattern between tools that have no official joint documentation. For teams evaluating whether a new agent tool or pattern is worth adopting, the MarkTechPost notebooks often appear before official tutorials do.
Maintenance and project status
The last push to the repository was on 2026-09-04. The repository is not archived. There are no GitHub releases, which means updates arrive as commits without version numbers or a formal changelog.
The repository carries no licence statement in the top-level entries. The README does not include a licence section. This is a gap worth noting for any team that wants to use the code in a commercial product, because the default copyright position without a licence is that the author retains all rights.
New notebooks appear regularly. The README shows entries from 2026-06 through 2026-09 in a single scroll, suggesting an active publication cadence. The breadth of topics continues to grow: the most recent categories include GPT-5, MiniMax, Mirascope and OAuth 2.1 for MCP Servers.
Editorial conclusion
AI-Agents-Projects-Tutorials is most useful for developers who learn by reading and running concrete code alongside explanatory articles. Each notebook links to its companion piece on marktechpost.com, which explains the design decisions. The repository does not provide a framework, a reusable library or installable package. It is a reference archive: browse by folder, find a technique or model you need, open the notebook and read the article together. New entries appear frequently, with the last push on 2026-09-04, but there are no formal releases and no versioned changelog. A developer who needs a stable, documented agent framework should look at LangGraph or AutoGen directly; this repository is useful for seeing how those frameworks are applied to concrete tasks.
Frequently asked questions
What should I learn to build AI agents?
This repository shows concrete implementations of tool calling, session memory, multi-agent coordination, planning and reasoning loops across frameworks like LangGraph, AutoGen, CrewAI and direct Anthropic API calls. Each notebook pairs with an article explaining the design. Starting with the AI Agents Codes/ and Agentic AI Codes/ folders gives a practical foundation.
What are AI agent projects?
This repository contains dozens of runnable AI agent projects covering tasks like financial research workflows, autonomous data science agents, web crawling pipelines, skill-driven Claude integrations and nanobot-style personal assistants. Each project is a Jupyter notebook or Python script paired with a tutorial article.
Where can I find tutorials for AI agents?
This repository links each notebook to a companion article on marktechpost.com that explains the design. The topics span LangGraph, AutoGen, CrewAI, direct Anthropic API usage, MCP connectors and many other frameworks and tools.
What are the 7 types of AI agents?
The README does not enumerate a taxonomy of AI agent types. The repository covers a variety of agent patterns including multi-agent systems, memory-equipped agents, planning agents and reasoning loop agents, but does not organize them into a canonical numbered list.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/marktechpost-ai-media-inc-ai-agents-projects-tutorials)