Better Agents CLI
A LangWatch command-line tool that creates the scaffold of a new AI agent project and hands it to a coding assistant
- 1,558 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The wizard accepts LangWatch and model provider keys and writes them to the project .env
- Launches the chosen coding assistant with an initial prompt and creates files in the chosen folder
- Sends anonymous usage telemetry by default
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/better-agentsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The wizard accepts LangWatch and model provider keys and writes them to the project .env
- Launches the chosen coding assistant with an initial prompt and creates files in the chosen folder
- Sends anonymous usage telemetry by default
This is third-party code. Review the repository files before installing.
What it does
Better Agents is a console tool from the LangWatch team for starting a new AI agent project. The `better-agents init` command asks for the programming language, agent framework, coding assistant, model provider and keys, then creates the project folder and launches the chosen assistant in it with a ready initial prompt. The project gets a fixed structure: agent code in app/ or src/, scenario tests and Jupyter evaluation notebooks in tests/, versioned prompts in prompts/ with a prompts.json registry, a .mcp.json file and development rules in AGENTS.md. The rules require every new feature to get a scenario test and every prompt change to be versioned. .mcp.json registers MCP servers with documentation for the chosen framework and for writing Scenario tests, so the assistant checks against current documentation.
Who it is for. For developers who start a new AI agent with Claude Code, Cursor or Kilocode on Agno, Mastra or LangGraph and want tests and prompt tracking from day one.
Good fit when
- You are starting a new agent and want the assistant to write scenario tests from the first feature
- A team needs one project layout for agents with versioned YAML prompts
- You are choosing among the supported frameworks and want the assistant to read the chosen one's documentation through MCP
- You need notebooks that evaluate parts of the pipeline, such as RAG or classification, with a dataset next to the code
Not a fit when
- A project is already under way: the tool creates a new scaffold and does not restructure an existing repository
- Your framework is not one of the three supported ones, so there is no template for it
- You need a CLI to run the agent in production: Better Agents only creates the project and launches the assistant
Example request
Use better-agents to create a new TypeScript agent project with Mastra and prepare a first scenario test for a customer asking to return an orderLimitations
- Node.js 22 or newer, npm or pnpm and one of the assistants are required: Claude Code, Cursor, Antigravity or Kilocode CLI. The CLI shows which of them are already installed.
- A LangWatch key (issued at app.langwatch.ai) and a key for the chosen model provider are required. OpenAI and Anthropic APIs are unavailable from Russia, so check your provider before installing.
- The template is tied to LangWatch: scenario tests, evaluations and prompt versioning go through its services and packages.
- The tool is at 0.x, so the scaffold format may change between releases.
How to disable. Remove the package with `npm uninstall -g @langwatch/better-agents`. A generated project stays an ordinary folder; AGENTS.md and .mcp.json can be edited or deleted by hand.
FAQ
How do I install Better Agents and create a project?
Globally with `npm install -g @langwatch/better-agents`, or without installing: `npx @langwatch/better-agents init my-agent-project`. The wizard then asks for language, framework, assistant, provider and keys.
Which frameworks and assistants does Better Agents support?
Agno, Mastra and LangGraph are supported. For Agno and Mastra the project gets an MCP server with documentation. The assistants are Claude Code, Cursor, Antigravity and Kilocode CLI, and the CLI detects which are installed.
Can I create the project in an existing folder?
Yes, `better-agents init .` creates the scaffold in the current folder and adds a .env file with keys and a .gitignore. The tool does not restructure existing code; it is built for starting a new project.
What is the agent testing pyramid in Better Agents?
Three levels. Unit tests cover deterministic code, notebook evaluations measure probabilistic parts such as RAG, and Scenario tests run the agent as a conversation partner from start to finish.
Does Better Agents collect usage data?
Anonymous telemetry is on by default. Turn it off with the `BETTER_AGENTS_TELEMETRY=0` environment variable.
Discussion
No questions or reviews yet.
Related
CLI toolsA multi-agent harness that verifies results by artifacts: stop hooks, an independent judge and an event log for Claude Code, Codex, Cursor and more
A team of AI agents from one request: it plans the task, writes code in parallel branches, tests, reviews and opens a pull request
A CLI to run several coding agents: one config, a session console, local memory and a set of engineering skills
A skill with 25 commands to audit and improve the prompts, tools and agent designs of a project in Cursor, Claude Code and Codex
An open coding agent for the terminal, browser and desktop that works with models from OpenAI and Claude to Ollama
A memory and coordination layer for coding agents: specs and issues live in git while agents work in parallel
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| Better Agents (this entry) | no | 1,558 | |
| project-starter | no | 1,420 | a universal set of agents, commands and hooks for Claude Code, with no project scaffold or agent tests |
| mcp-use | no | 10,828 | a plugin for building MCP servers in TypeScript, with no AI agent project scaffold |
README in short
The idea: a tool and a set of standards make the coding assistant an expert in the chosen framework. Details live in separate documents: quick start, walkthrough, project structure and features.