The goal of langgraph-hierarchies is to make deep, stateful agent hierarchies on LangGraph decomposable. Flat supervisor handoffs and harness-style subagents are a good starting point; this library targets the wall beyond that — recursive trees where subagents are real compiled subgraphs, with explicit policy for how parent and child state interact when a subchain starts, runs, and exits.

In summary, langgraph-hierarchies is an open-source Python library for production-grade nested agents: declarative state isolation, supervisor-controlled iteration limits, and phased compilation — not ephemeral task-tool isolation alone.

I'm building it around:

• Class-as-factory graphs with phased compile (BaseGraph / CompiledGraph, root compile via compile_as_root)
• SubchainPolicy — entry snapshot, per-field clear / merge / discard, exit restore
• ReactGraph with iteration safety (per-agent limits, supervisor task_iterations, forced-exit reporting)
• TodoGraph + todo toolkit, with an IRS reporting hierarchy example in the repo

This project is particularly interesting to me because it generalises problems I've hit building multi-agent systems in production — especially complex agent hierarchy, accountability across nested runs, and keeping subgraph state predictable when depth goes past one supervisor layer. Early development (0.0.x); v0.1 will ship the keystone slice. Not affiliated with LangChain or the LangGraph team.

If you're building LangGraph agents that need more than flat delegation — nested teams, isolated subchains, and explicit state policy — you can install it from PyPI or browse the source and examples on GitHub.

Also available on PyPI: pip install langgraph-hierarchies

[ Python ] [ LangGraph ] [ uv ] [ pytest ] [ ruff ] [ GHA ]

Navigator is an application that aims at providing high-quality pentesting service for companies that can't purchase a specialised penetration tests. By automating selected section of the OWASP Top 10 (2025) classification we reduce costs and provide our customers with complex support, guiding them from detection of vulnerabilities and weaknesses in application design (OWASP A06: Insecure Design) verifying their deployment pipeline (OWASP A08: Software or Data Integrity Failures), examining the attack surface for most common vulnerabilities through explaining the nature of vulnerabilities and offering guidance in fixing and re-testing them.

In summary, Navigator offers affordable autonomous pentest teams at fraction of price and with expanded capabilities compared to traditional scanners.

This project is for me particularly interesting because I have an opportunity to tackle such operational issues as:

• complex agent hierarchy
• model and application observability
• multi-tenancy in multi-agent application
• provisioning and pricing
• figuring out early adoption

You can see Navigator in action here:

Also: here and here

[ Python ] [ TypeScript ] [ LangChain ] [ LangGraph ] [ LangSmith ] [ BrowserUse ] [ OpenHands ] [ Prometheus/Grafana ]
[ AWS ] [ openrouter.ai ] [ HF TGI ] [ HF TRL/SFT ] [ GHA ] [ Terraform ] [ Docker ]

The goal of the Toshokan project is to aid English-speaking students in learning Japanese language. In my experience on the path to learn Japanese I realised that there are many obstacles that I haven't come across learning e.g., English or Portuguese. For example:

• Formality levels differ depending on age & hierarchy
• Exposure to language is problematic. Popular multimedia don't contain equivalent of everyday Japanese
• Kanji logographic register is very large and time-consuming

‍

‍

In Toshokan I attempted to harness capabilities of contemporary LLMs to remove these barriers. I'm doing so by:

• Controlling the list of known kanji and kanji that you want to schedule to memorise
• Selecting formality level for practicing different language registers. Using LLMs capability to properly select language register for formal / semi-formal / informal situations
• Breaking down the learning process into individual lesson units with limited scope that you can exercise ad infinitum or until you are satisfied
• Including a fast chat for words and auxiliary chat for general questions in context of Japanese
• Listing and annotating all unknown kanji (kanji outside of the known kanji register) during practice conversation

If you're an Japanase N4-N2 level student, I think you might find it useful. Feel free to check it out here and let me know your thoughts!

[ Python ] [ TypeScript ] [ LangChain ] [ LangSmith ] [ LangGraph ] [ openrouter.ai ] [ Supabase ]
[ Render.com ] [ Vercel ] [ GHA ] [ Docker ]