Report Abuse

Basic Information

Fabrice AI is a lightweight TypeScript framework for building composable AI agents that collaborate to solve multi-step tasks. It provides primitives to define agents, workflows, tools and providers, and to run coordinated execution in a functional, stateless style. The project is infrastructure-agnostic and designed to be serverless-ready, so teams can plug in different LLM providers or run workflows locally or in the cloud. The library supplies helpers for creating initial, root and child workflow states and exposes a simple execution loop so developers can run, pause, resume, or manage long-running operations. It also includes a project creator utility to scaffold examples and templates.

Links

Categorization

App Details

Features
Functional, stateless API for agents and workflows implemented in TypeScript. Built-in orchestration helpers such as teamwork and iterate to execute workflows to completion or in paused, resumable cycles. Batteries-included set of built-in agents (supervisor, resourcePlanner, finalBoss), built-in providers for OpenAI and Groq, and an extensible provider interface so custom providers can be implemented. Tooling support via a @fabrice-ai/tools package and an API for creating custom tools with parameter schemas. Helpers for message creation, delegating and handing off tasks, and managing hierarchical workflow state. A small CLI scaffolder create-fabrice-ai and a bundled BDD testing package for behavior-driven tests.
Use Cases
Fabrice AI helps developers orchestrate multiple specialized agents without heavy object-oriented frameworks or infrastructure lock-in. Its composable, function-first approach reduces side effects and simplifies reasoning about state, which is beneficial for serverless and cloud deployments. Teams can replace built-in agents or swap providers per-agent, enabling experiments with different LLMs and tool integrations. The execution model supports pausing for manual tool execution and resuming, which is useful for long-running tasks and cloud workflows. The included tools, provider adapters, scaffolding and BDD testing support accelerate building, testing and iterating on multi-agent workflows.

Please fill the required fields*