AxOpen-source TypeScript framework for building LLM agents and agentic workflows.

4.4 (5)

Reviewed by Daniel Nikulshyn·Updated May 2026

Open Source TypeScript LLM Framework AI Agents Developer Tools Node.js Self-Hostable

Overview

Ax is a TypeScript framework that helps developers build applications powered by large language models, with a focus on agents and multi-step workflows. It provides structured primitives for prompts, tools, and reasoning chains so teams can move beyond ad-hoc prompt engineering toward more maintainable, typed code. The framework draws on ideas from prompt-programming research, letting developers describe inputs and outputs declaratively while Ax handles formatting, parsing, and tool orchestration. It supports multiple LLM providers and integrates with existing Node.js and TypeScript stacks. As an open-source project, Ax is suited for engineers who want transparency, extensibility, and the ability to self-host their agent infrastructure rather than depend on a closed platform.

Key features

Declarative prompt and signature definitions
Agent and tool orchestration primitives
Multi-provider LLM support
Streaming and structured output parsing
Built-in support for chained reasoning
Extensible with custom tools and adapters

Pricing

Model: Freemium
Category: AI Agents
Rating: 4.4 / 5 (5)

Use cases

Build Typed LLM Agents

Create AI agents with structured inputs, outputs, and tool orchestration using TypeScript signatures instead of brittle ad-hoc prompt strings.

Multi-Step Agentic Workflows

Compose chained reasoning steps and tool calls into maintainable workflows for tasks like research, data extraction, or automated decision-making.

Provider-Agnostic LLM Apps

Develop applications that work across multiple LLM providers, allowing teams to switch or compare models without rewriting core logic.

Self-Hosted AI Integrations

Embed open-source LLM capabilities into existing Node.js and TypeScript stacks where transparency, extensibility, and self-hosting are required.

Pros & Cons

Pros

Strong TypeScript typing and developer ergonomics
Open-source and self-hostable
Works with multiple LLM providers
Structured approach to agents and tools

Cons

Requires TypeScript and coding knowledge
Smaller ecosystem than larger frameworks
Documentation can lag behind rapid changes

Reviews

4.4

Average from 5 ratings.

Margaret Whitfield

Feb 27, 2026

Years in this space

I've evaluated a lot of these over the years. What stands out here is agent and tool orchestration primitives — handled better than most — and strong TypeScript typing and developer ergonomics. Requires TypeScript and coding knowledge is my one real gripe. Worth the time if this is your use case.

Aaliyah Johnson

Nov 10, 2025

Does the job

Pretty happy overall. Agent and tool orchestration primitives just works and open-source and self-hostable. Documentation can lag behind rapid changes can be annoying, but no dealbreakers — I'd recommend it to a friend without hesitating.

Pierre Dubois

Nov 9, 2025

Compared a few options

Evaluated this against two competitors. Where it wins: declarative prompt and signature definitions and strong TypeScript typing and developer ergonomics. On balance the feature set — especially agent and tool orchestration primitives — justifies the 5 stars for our use case.

Wei Chen

Sep 21, 2025

Years in this space

I've evaluated a lot of these over the years. What stands out here is extensible with custom tools and adapters — handled better than most — and strong TypeScript typing and developer ergonomics. Requires TypeScript and coding knowledge is my one real gripe. Worth the time if this is your use case.

Carlos Mendoza

Aug 1, 2025

Use it every day

Honestly didn't expect to like it this much. Declarative prompt and signature definitions is exactly what I needed, and structured approach to agents and tools. I do wish requires TypeScript and coding knowledge, but I reach for it almost every day now and it just clicks.