Langflow

Langflow is an innovative technology provider specializing in the integration and interaction with language models. Langflow's solutions facilitate effortless connection to various language models, enabling powerful and intuitive conversational interfaces.

General
Author	Langflow
Repository	https://github.com/langflow
Documentation	https://docs.langflow.org/
Type	Language Model Integration Platform

Key Features

Provides robust APIs for easy integration with multiple language models, enhancing conversational applications
Delivers high performance and scalable solutions to manage conversational workflows
Simplifies development of language-driven applications with a minimal configuration requirement
Ensures efficient handling of multiple simultaneous conversations, maintaining performance as usage scales

Start building with Langflow's products

The Langflow API enables developers to easily connect to and manage language models, supporting a range of functionalities from basic querying to complex conversational interactions. The API is designed to be intuitive and developer-friendly, allowing for quick integration and robust support for diverse application needs.

List of Langflow's products

Langflow API

Langflow Studio

Langflow Studio provides a comprehensive environment for designing, testing, and deploying language model interactions. The studio's user-friendly interface allows developers to visually construct dialog flows and fine-tune responses, ensuring that applications deliver natural and effective user interactions.

Langflow Hub

Langflow Hub serves as a central repository for pre-built language model templates and configuration presets. It offers developers a quick start to building applications with pre-configured setups for common use cases, from customer service bots to interactive educational guides.

Starter Projects

Basic Prompting

Prompts are inputs for a large language model (LLM), bridging human instructions and computational tasks. Enter natural language requests in a prompt to get answers, generate text, and solve problems.

Blog Writer

The blog writer leverages dynamic, URL-based references to ensure the content is accurate and relevant. Use Langflow to build a blog writer with OpenAI that utilizes URLs for reference content.

Document QA

Build a question-and-answer chatbot with a document loaded from local memory.

Memory Chatbot

Extend the basic prompting flow to include chat memory for unique SessionIDs.

Vector Store RAG

Retrieval Augmented Generation (RAG) is a method for training large language models (LLMs) on the specific dataset and querying it effectively. It utilizes a vector store to store embeddings of the data, enabling advanced and context-aware search capabilities.

System Requirements

Langflow is compatible with Linux, macOS, and Windows operating systems, requiring at least 4 GB of RAM and adequate storage for development data. A multicore processor is recommended to handle multiple requests efficiently, with a stable internet connection necessary for accessing cloud-based features. Modern web browsers with JavaScript enabled are required, while the use of GPU acceleration is optional but beneficial for optimizing performance.

quant-agents: Autonomous Quant Research on AMD

quant-agents is a local "Quant Hedge Fund in a Box": a multi-agent system that runs the full quant research loop end to end with zero cloud dependency. A LangGraph state machine routes work between four agents. The Ideator proposes a market hypothesis with an economic rationale, precise statistical entry and exit rules, and regime filters. The Coder turns it into a fully vectorized pandas/numpy generate_signals function (no loops, no iterrows). A sandbox executes the code on SPY daily data, lags every signal with .shift(1), and computes Sharpe, max drawdown, total return, and win rate. The Auditor acts as head of risk: it diagnoses tracebacks, rejects lookahead bias and curve-fitting, and demands structural fixes like regime filters or volatility-based sizing instead of parameter tweaking. A strategy only passes with Sharpe above 1.5 and drawdown under 15%, checked in code, not by the LLM. Survivors go to the Packager, which writes a production handoff report with documented code, Big-O analysis, execution edge cases, and a C++ translation plan (Eigen/Armadillo). The whole system runs on a single Qwen3.6-35B-A3B Mixture-of-Experts model served by vLLM on AMD Instinct MI300X with ROCm. Instead of loading four models and fragmenting HBM, all agents share one model and LangGraph swaps the persona per call. With only ~3B active parameters per token, the agent loop stays fast enough to iterate many times per strategy. Circuit breakers cap rework attempts and new ideas, so the system either finds a strategy that survives the audit or stops cleanly.

VoiceForge AI – Smart Voice Assistant

VoiceForge AI is an agentic productivity assistant that provides a real-time voice interface for interacting with documents, tasks, calendars, application data, and external tools. Users can speak naturally, ask questions about uploaded PDF, Markdown, text, CSV, and JSON documents, create and manage tasks, review calendar events, query application data, and interact with integrations such as GitHub, email, and REST APIs. The system uses AssemblyAI Voice Agent API for live speech recognition, audio responses, and voice sessions, while Gemini provides intelligent reasoning, routing, structured decisions, tool usage, and answer verification. LangGraph orchestrates specialist agents for knowledge, research, tasks, data, and actions. Qdrant enables user-scoped document retrieval, while MongoDB stores users, sessions, messages, tasks, documents, and audit data. Security is integrated throughout the platform using JWT authentication, permissions, rate limiting, request validation, audit logs, secure backend credentials, and human approval for sensitive actions. The frontend is built with React, TypeScript, Vite, Tailwind CSS, and Web Audio API, while the backend uses Python, FastAPI, Pydantic, LangGraph, LangChain, and WebSockets. VoiceForge AI brings voice AI, multi-agent orchestration, RAG-based document retrieval, productivity automation, and secure tool execution together in one unified platform.

PitchPI-AI_B2B_Sales_Agent

The Real-Time AI B2B Sales Agent is a comprehensive solution engineered to revolutionize the business-to-business sales process. By leveraging advanced artificial intelligence and agentic dialogue workflows, this application acts as an autonomous sales representative capable of conducting human-like, real-time voice conversations with prospective clients. The system's architecture is highly robust, utilizing a Python-based dialogue engine that integrates Speech-to-Text (STT) and Text-to-Speech (TTS) services for seamless, low-latency voice interaction. To ensure the agent provides accurate, context-aware responses, it employs a Retrieval-Augmented Generation (RAG) pipeline powered by Chroma. This allows the agent to pull specific product information and company knowledge on the fly during a call. Furthermore, the project features a highly scalable backend infrastructure. It uses Redis for high-speed caching and MongoDB for persistent data storage. A dedicated Java Spring Boot microservice seamlessly synchronizes lead data, interaction logs, and call analytics via REST APIs. We also incorporated K-Means clustering algorithms for intelligent B2B lead enrichment and scoring. With a responsive frontend deployed on Netlify, this project represents the next generation of scalable, automated communication tools for modern sales teams.

Clocket AI – Smart Developer Onboarding Assistant

## Problem Joining an unfamiliar software project can take hours or even days. Developers must manually understand the repository structure, technology stack, architecture, dependencies, environment configuration, APIs, database usage, and contribution workflow. README files may be incomplete or outdated, making the onboarding process slow and difficult, especially for new developers. ## Solution Clocket AI is a Smart Developer Onboarding Assistant that automatically analyzes a public GitHub repository and converts the codebase into an interactive onboarding experience. A developer signs in, provides a GitHub repository URL, and starts an automated analysis. Clocket AI validates and scans the repository, detects its technologies and configuration, analyzes the architecture, generates an evidence-based setup guide, creates practical starter tasks, indexes relevant source files into a repository-scoped RAG system, and stores the results for the user. Developers can then explore the generated architecture, follow setup instructions, work through beginner/intermediate/advanced starter tasks, track their onboarding progress, and ask natural-language questions about the codebase. Answers are grounded in retrieved repository files and include source citations. ## Target Users The primary users are developers joining unfamiliar projects, interns, students learning existing codebases, engineering teams onboarding new members, and maintainers who want to make repositories easier to understand. ## Technology and Architecture The application uses a React/Vite frontend and FastAPI backend. LangGraph orchestrates a multi-stage analysis workflow including repository validation, cloning, deterministic static scanning, architecture analysis, setup-guide generation, onboarding-plan generation, task generation, RAG indexing, persistence, and cleanup. The system combines deterministic analysis with LLM reasoning rather than relying entirely on an LLM. Repository files are chunked and embedded using fastembed and stored in Supabase PostgreSQL with pgvector. Retrieval is filtered by both user and repository to maintain data isolation. Clocket AI also integrates the official GitHub MCP Server for repository access and custom MCP tools for saving analysis results and managing onboarding progress. ## What Makes It Unique Instead of simply generating a README or chatbot response, Clocket AI creates a complete, interactive onboarding workflow from the actual codebase. It combines repository analysis, Agentic RAG, MCP tools, architecture explanation, evidence-based setup instructions, actionable starter tasks, and progress tracking in one system. Security is also built into the architecture through Supabase authentication, Row Level Security, repository-scoped retrieval, server-side secrets, read-only GitHub MCP access, file-size limits, and prompt-injection-resistant handling of repository content. The application is deployed and available as a web application, making the onboarding workflow accessible without requiring developers to manually configure the system locally.

Conversational Sales Agent for Enterprise Websites

Modern B2B websites lose up to 70% of potential inbound leads because passive documentation, static contact forms, and slow, text-only chat widgets create friction. Prospective buyers want immediate answers to high-stakes queries , such as service capabilities, technical stack fit, and project scoping , without waiting 24 hours for a sales representative to reply. VoiceFlow (featuring QM Assistant as our live enterprise case study) is an ultra-low-latency, voice-driven conversational sales agent engineered to bridge this gap. By combining AssemblyAI’s Streaming Speech-to-Text, Groq LPU inference, and real-time audio synthesis, VoiceFlow acts as an always-on digital account executive that converses naturally with website visitors, protects enterprise pricing boundaries, and autonomously triggers high-converting sales actions.

Taible OrderFlow

Taible Voice Orchestration is the CPU turn-taking layer for the Taible voice agent, built as a Pipecat/LiveKit Agents worker. It bridges a diner's browser microphone to a fully hosted STT to LLM to TTS pipeline and drives ordering through the Taible MCP server, requiring no GPU. Audio flows from the browser mic over WebRTC into a LiveKit room, through the orchestrator, and back out to the browser speaker. Inside the orchestrator, Deepgram handles speech-to-text as the ears, Fireworks AI's gpt-oss-120b model serves as the brain, and Deepgram again handles text-to-speech as the mouth. MCP tools connect out to a Cloud Run service backed by Supabase, and the agent writes live transcript and cart state to the sibling Next.js UI's public folder. The full voice loop has been verified end-to-end in Docker: a diner speaks, the AI waiter replies with voice, updates the cart, and persists the order. On each session, the agent bootstraps against the MCP server to resolve the restaurant and menu, starts a session and order, and adds items as the diner orders them, resolving spoken names to real menu-item IDs. Order confirmation is intentionally left to the diner, who taps a Confirm Order button in the UI rather than letting the AI place it. The stack is configurable via environment variables for LiveKit, Deepgram, Fireworks, and the restaurant/table identifiers, and can run either via Docker Compose or a local Python virtual environment alongside the UI.