langchain

microsoft/markitdown

Turning office files and media into LLM-ready Markdown — MarkItDown converts PDFs, Office documents, images, audio, HTML and more into Markdown, aimed at feeding text analysis pipelines and LLMs rather than human-facing output. It's comparable to textract but focuses on preserving structure like headings, lists, tables and links. Useful for anyone building retrieval or document pipelines with OpenAI or Autogen.

Python tool for converting files and office documents to Markdown.

bytedance/deer-flow

DeerFlow 2.0: a ground-up rewrite of ByteDance's super agent harness — DeerFlow 2.0 is a complete rewrite orchestrating sub-agents, memory, sandboxes and extensible skills to handle long tasks from minutes to hours. It recently hit #1 on GitHub Trending. Useful for developers building agentic research, coding, and content-creation workflows with LangGraph/LangChain.

An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message gateway, it handles different levels of tasks that could take minutes to hours.

headroomlabs-ai/headroom

Shrinking AI agent context without losing the important lines — Headroom compresses tool outputs, logs, and RAG chunks before they hit the LLM, cutting tokens by 20% for coding agents and up to 95% for JSON while preserving exact answers, including critical lines like error messages. It ships as a Python/TS library, proxy, or MCP server, with its own kompress-v2 model on Hugging Face.

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.

BerriAI/litellm

LiteLLM moves to a Rust core for a leaner AI gateway — LiteLLM now ships a Rust core under its Python SDK, aiming to be a faster, lighter self-hosted AI gateway. It unifies 100+ LLM APIs (Bedrock, Azure, OpenAI, Anthropic, VertexAI, vLLM, Nvidia NIM) in OpenAI format, adding cost tracking, guardrails, load balancing and logging for teams running LLMOps at scale.

The fastest, litest AI Gateway. Rust core with Python SDK. Call 100+ LLM APIs in OpenAI (or native) format with cost tracking, guardrails, load balancing, and logging [Bedrock, Azure, OpenAI, Anthropic, OpenAI, VertexAI, vLLM, Nvidia NIM]

langfuse/langfuse

Open-source tracing and evals for LLM apps and agents — Langfuse is a self-hostable platform to trace, evaluate, and debug LLM applications, with integrations for LangChain, LlamaIndex, and AutoGen plus a prompt playground and management tools. Backed by Y Combinator, it suits teams needing open, self-hosted observability instead of closed SaaS-only tools.

🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.

1Panel-dev/MaxKB

An enterprise agent platform combining RAG, workflows and MCP tools — MaxKB pairs retrieval-augmented generation with a workflow engine and MCP tool-use, letting teams build knowledge-grounded agents for customer service, internal knowledge bases, and education. It supports document upload, web crawling, and models like DeepSeek-R1, Llama3 and Qwen3, aimed at enterprises wanting lower hallucination Q&A.

🔥 MaxKB is an open-source platform for building enterprise-grade agents. 强大易用的开源企业级智能体平台。

thesysdev/openui

A framework-agnostic language for streaming AI-generated UI — OpenUI defines a compact, streaming-first language for generative UI, with runtimes for React, Vue, Svelte, and Angular. It claims up to 67% fewer tokens than JSON, and React gets built-in component libraries plus ready-made chat interfaces. Aimed at developers building LLM agent frontends.

The Open Standard for Generative UI

cloudwego/eino

A Go-native framework for LLM apps and agents, built by CloudWeGo — Eino is a Golang LLM application framework inspired by LangChain and Google ADK but built to Go conventions. It offers reusable components (ChatModel, Tool, Retriever), an Agent Development Kit for multi-agent coordination and human-in-the-loop interrupts, and graph-based composition for workflows. Aimed at Go developers building production AI agents without leaving the Go ecosystem.

The ultimate LLM/AI application development framework in Go.

BitterSecurity/Decepticon

An autonomous agent framework for red team pentesting — Decepticon is an open-source autonomous agent built on LangChain/LangGraph that performs red team tasks beyond running nmap and writing reports. Trending on Trendshift, it targets security researchers and pentesters wanting LLM-driven, agentic exploit testing rather than scripted scanners.

Autonomous Hacking Agent for Red Team

strelov1/freehire

3.3M IT jobs scraped straight from company career pages — freehire is a self-hosted, open-source job search engine that crawls ATS platforms like Workday, Greenhouse, Lever and Ashby directly, skipping recruiter reposts and dead links. Built with Go, SvelteKit, Meilisearch and Postgres, it targets IT job seekers and anyone wanting a transparent, self-hostable alternative to closed job boards.

freehire — the open-source search engine for job seekers