Report Abuse

Basic Information

Project Alice is an open-source agentic workflow framework designed to create, manage and deploy AI agents and automated tasks. It combines conversational chat capabilities with a task execution engine in a microservices architecture that uses MongoDB for persistence. The repository contains three main components: a Node.js/Express TypeScript backend that manages the database, LM Studio interactions and file serving; a Python Workflow container built with Pydantic that implements task and chat execution logic; and a React TypeScript frontend that provides a flow viewer, task editor and chat UI. The project is model-agnostic and supports local and remote models, local LM Studio deployments and many API providers. It targets developers and teams who want a flexible environment to prototype agent workflows, run retrieval-augmented workflows, and deploy agents programmatically via exposed workflow endpoints.

Links

Categorization

App Details

Features
Alice bundles task-oriented workflow execution, an intelligent chat system and extensible agent models. It supports Retrieval-Augmented Generation via a RetrievalTask and embeddable Data Clusters, Human-in-the-loop checkpoints for pausing tasks, a basic chain-of-thought mechanism, and AliceDocuments for structured system prompts. Node-based task flows allow complex routing between inner nodes and workflows that call other tasks. The frontend includes a flow viewer and prompt parser, and the Workflow container exposes APIs such as /execute_task and /chat_response/{chat_id}. The platform supports many API types and providers (LLM models, embeddings, search, Wolfram Alpha, Google Knowledge Graph, PixArt, Bark TTS and more), local LM Studio models, text splitters and message pruning, a Redis queue for concurrent execution, threads in chats and role-based user/admin management.
Use Cases
This repository helps developers and teams build, test and deploy agentic solutions with a reusable framework that separates UI, orchestration and execution. It simplifies creating custom tasks (API tasks, prompt agent tasks, code execution and generation tasks, web scraping, TTS, image generation, embedding and retrieval), visualizing task logic and inspecting node-level results. Programmatic access to the Workflow API enables automation and integration into other systems. Built-in features like data clusters, local model support, cost and context tracking, user checkpoints and admin functions make it practical to iterate on workflows, manage access and run local or remote models. Docker-based setup and update scripts enable local deployment and experimentation while keeping the project model- and provider-agnostic.

Please fill the required fields*