# Phoenix Experiment Runner
This is an example app that uses the phoenix-client and @clack/prompts to run experiments.
It runs against the `llama3.2` model, pointing to a local ollama instance.
## Setup
### Requirements
- [Phoenix](https://phoenix.arize.com/) running locally on `http://localhost:6006`
- `pnpm d:up` will start Phoenix
- Create a dataset from the `qa-dataset.csv` file in this directory in Phoenix at `http://localhost:6006/datasets`
- [Ollama](https://ollama.com/) with `llama3.2` installed OR OpenAI credentials configured in `.env`
- You may copy the `.env.example` file to `.env` and fill in the values if you do not wish to use Ollama
- Node.js 20+
- pnpm
### Install dependencies
```bash
pnpm install
pnpm -r build
```
This will build all the dependencies in the monorepo and the app.
### Run the app
```bash
pnpm d:up
pnpm dev
```
This will start Phoenix at `http://localhost:6006` and the app in dev mode.
### Build the app
```bash
pnpm d:up
pnpm build
pnpm start
```
This will start Phoenix and build the app, then run the app.
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Arize-ai/phoenix'
If you have feedback or need assistance with the MCP directory API, please join our Discord server