amitshekhariitbhu/ai-agents-tutorial
Learn AI Agents step by step, from scratch - from function calling to agent loops to multi-agent systems, orchestration, and evaluation.
Topic: tool-calling-agent
37 repositories
TheUncharted/zapcode
TypeScript interpreter for AI agents. Written in Rust. 2µs cold start. Sandboxed. Alternative to MCP tool calling.
decisionnode/DecisionNode
CLI + Local MCP - A shared structured memory store across Claude Code, Cursor, Windsurf, Antigravity, and every MCP client. Semantically queryable.
PushpenderIndia/browsegenie
AI Agent which can do any kind of Browser Automation Task & Web Scraping, just using a single prompt!
pguso/voice-agents-from-scratch
From-scratch voice agents in Python: end-to-end speech pipelines, runnable chapters, and a small shared library. Local models, explicit streaming behavior.
CaptCanadaMan/pi-ollama
First-class local model support for pi through Ollama — reliable tools, context management, thinking controls, vision, warm-up, and telemetry.
sx4im/resultseal
Deterministic observation-integrity contracts for AI-agent tool results. HTTP 200 is not an observation.
masteranime/n8n-rails
Force deterministic tool execution order in n8n AI Agents. No skipping, no hallucinated parameters.
yiyangzhang-ai/open-agent-eval
Lightweight open-source toolkit for evaluating tool-calling AI agents on safety, correctness, and reliability
Lonishubh48/insight-link-pro
Hallucination Killer — MCP server that grounds LLM answers in live repo context and real-time docs
coquencio/LiteAgent
A high-performance, token-efficient AI Agent tool-calling library for .NET. Replaces JSON with TOON for token savings.
viochris/NovaCal-AI-Streamlit
An advanced AI scheduling assistant powered by Gemini 2.5 Flash and Streamlit. Effortlessly manage your Google Calendar using natural language and a visual dashboard.
24prathamesh2004/pulse_blog_agent-POD-A
Pulse is an autonomous newsroom that discovers, curates, and publishes content entirely on its own. No human writers needed—just configure your editorial policy and let the AI agents do the rest.
ChandralekhaSelvam/slm-toolcall-lora
QLoRA fine-tuning of Llama 3.2 3B for reliable structured tool-call generation. Investigating whether ~3,000 curated examples can close the reliability gap between small open models and hosted APIs (GPT-4o, Claude) at <5% of the cost. Runs on free Colab. Data quality > quantity.
apex-freen/fde-mcp-blade
The AI toolbox you deploy inside your customer's data center — one unified MCP entry with five-layer security and full audit trails. Plugin engine turns Feishu / DingTalk / WeCom / intranet APIs into agent tools. One Docker command, 1-core/1GB min, data never leaves your network. Apache 2.0.
lapheavisidelayer940/DecisionNode
Store and query shared structured memory across Claude Code, Cursor, Windsurf, and any MCP client with a local CLI and MCP server
brunogcar/agent
Local-first Cognitive Execution Platform for Multi-LLM Orchestration, Autonomous Workflows, Persistent Memory & Offline Learning
SC-LABS-ai/sc-node
Experimental low-overhead Rust agent harness for running tool-using agents across local and cloud LLM providers.
Jahnavi-Krishna/aria--smart-hvac-parts-agent
A grounded AI parts agent for HVAC technicians — GPT-4o tool-calling, deterministic compatibility checks (never guessed), photo-based part lookup, and voice replies.
prakharpd/Agentic-Portfolio
An AI chatbot with tool calling that answers questions about your professional portfolio using AI Agents and RAG running locally on your PC for 24/7. It can also capture leads, answer questions, showcase expertise automatically. Deployed Link (NO TOOL CALLING !, NO CONTACT CAPTURE !)👇🏻
Jahnavi-Krishna/roofintel--b2b-lead-agent
Mini B2B sales intelligence agent with explainable scoring, GPT-4o insights, and real actions (drafts outreach, updates records).
Massi99RM/My-First-Agent
A minimal ReAct agent built with LangChain and LangGraph to understand how LLM agents work from first principles.
jihadkhawaja/OneShotPrompt
one prompt, one shot, done.
tinyinternetclub/consentry
A local policy gate for AI agent tool calls.
Karan-Negi-12/Project-Jarvis
JARVIS is a modular AI voice assistant powered by Gemini, Faster-Whisper, and Python. It features wake-word detection, speaker verification, offline STT/TTS, memory, web search, automation tools, and an extensible agent architecture inspired by Iron Man's JARVIS.
charlesericwilson-portfolio/Echo_Project_Overview
Overall description of multiple linked iterations of LLM agent frameworks.
Yasin1ar/dir-manager-agent
single-agent, tool-calling system, for managing directory files.
jiugao-yi/xihe-agent
Xihe（羲和）— 单进程 Tool-Calling Agent，一个大脑三种形态：CLI、企微/飞书网关、桌面应用。内置真实浏览器自动化、桌面操控、危险操作审批门控、专家子代理、业务知识库。
serey-roth/aml-investigation
AI-powered AML case management tool with a LangGraph agent investigating flagged patterns using past evidence and recommending SAR filing or closure.
habibun/symfony-ai-agent
A Symfony AI agent using OpenRouter, Doctrine-backed conversation memory, and tools for notes, web search, and email.