Arize Phoenix MCP
AIhttp-streamingstdio
Arize AI's Phoenix, an open-source AI observability and evaluation platform, exposes a remote MCP endpoint that connects agents directly to a Phoenix instance. Lets assistants query OpenTelemetry traces and spans from LLM applications, inspect datasets and experiments, review evaluation results, and investigate latency or quality regressions.
Подключение
Добавьте эту конфигурацию в .claude/mcp.json
Подключение:
{
"mcpServers": {
"arize-phoenix": {
"url": "https://github.com/Arize-ai/phoenix",
"env": {
"PHOENIX_COLLECTOR_ENDPOINT": "<YOUR_PHOENIX_COLLECTOR_ENDPOINT>",
"PHOENIX_API_KEY": "<YOUR_PHOENIX_API_KEY>"
}
}
}
}Инструменты (2)
query_spans
Query spans and traces recorded from an LLM application.
{
"type": "object",
"required": [
"project_name"
],
"properties": {
"project_name": {
"type": "string",
"description": "Phoenix project name"
},
"filter_condition": {
"type": "string",
"description": "Optional filter expression over span attributes"
}
}
}list_experiments
List experiments and their evaluation results for a dataset.
{
"type": "object",
"required": [
"dataset_name"
],
"properties": {
"dataset_name": {
"type": "string",
"description": "Dataset name"
}
}
}Ресурсы
Этот сервер не предоставляет ресурсы.
Промпты
Этот сервер не предоставляет промпты.
Сведения о сервере
- Автор
- Arize AI (@Arize-ai)
- Репозиторий
- https://github.com/Arize-ai/phoenix
- Лицензия
- Elastic-2.0
- Звёзды
- 0
- Последнее обновление
- 3 августа 2026 г.