# parseable-mcp-server [Health: Active]

**Category:** 📊 Monitoring  
**Repository:** https://github.com/parseablehq/parseable-mcp-server  
**GitHub Stars:** 3  
**npm Downloads (last month):** 703  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/parseable-mcp-server

## Description
Model Context Protocol server for Parseable. Lets LLMs discover and query Parseable.

## Tools
Capabilities this server exposes over MCP:

- **list_datasets** — List all log datasets
- **get_dataset_schema** — Column names + types
- **get_dataset_info** — Metadata (created_at, retention, time window)
- **get_dataset_stats** — Event count and storage bytes
- **sample_events** — Most recent N events (time-bounded, row-capped)
- **query_sql** — SQL `SELECT` over a time window. DDL/DML blocked. Auto-injects `LIMIT`.
- **query_promql** — PromQL instant or range query against a metrics dataset
- **list_alerts** — List all alerts with state, severity, tags
- **get_alert** — Full config for one alert
- **list_alert_tags** — All alert tags in use
- **enable_alert** — Enable an alert
- **disable_alert** — Disable an alert
- **evaluate_alert** — Force-evaluate now. **May fire real notifications.
- **create_alert** — Create alert via guided Q&A (8 questions, confirms before submit)
- **list_alert_targets** — List targets (Slack, webhook, Alertmanager)
- **get_alert_target** — Full config for one target
- **create_alert_target** — Create a new Slack / webhook / Alertmanager target
- **ping** — Check connectivity, return version + health
- **explain_query** — EXPLAIN` a SQL query without executing it
- **list_users** — List all users
- **get_user_roles** — Roles for a specific user
- **list_roles** — All role names
- **get_role** — Privilege definition for a role
- **get_default_role** — Default role for new users
- **get_cluster_status** — All nodes with status (distributed mode)
- **get_cluster_metrics** — Aggregated ingest/query/storage metrics
- **get_retention** — Retention policy for a dataset

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "parseable-mcp-server": {
    "command": "npx",
    "args": ["-y","@parseable/parseable-mcp-server"],
    "env": {
      "PARSEABLE_URL": "",
      "PARSEABLE_API_KEY": ""
    }
  }
}
```

**Requires environment variables:** `PARSEABLE_URL`, `PARSEABLE_API_KEY` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What parseable-mcp-server MCP server does

The parseable-mcp-server MCP server exposes Parseable operations as tools that an MCP-capable client can call. It can enumerate log datasets, inspect schemas and metadata, report event counts and storage usage, and return recent sample events. Dataset queries support SQL SELECT statements over a time window and PromQL instant or range queries for metrics datasets.

Alert workflows cover listing alerts and tags, retrieving an alert configuration, enabling or disabling alerts, and creating alerts through a guided eight-question flow with confirmation before submission. The server can also list and inspect Slack, webhook, and Alertmanager targets, or create a new target. The evaluate_alert operation forces an evaluation and may send real notifications, so it should be used with care.

Additional tools provide connectivity and query diagnostics, including ping and SQL EXPLAIN. Read-only RBAC tools expose users, user roles, role names, privileges, and the default role. In distributed Parseable deployments, administrative tools report cluster status and aggregate ingest, query, and storage metrics; another tool returns a dataset's retention policy.

## How it works

The parseable-mcp-server MCP server supports two transports. In stdio mode, the MCP client starts the process and the server reads Parseable connection settings from environment variables. In HTTP mode, one process serves a setup page and the Streamable HTTP MCP endpoint. Requests authenticate with X-API-Key. Cloud requests identify cloud mode, while self-hosted requests provide the Parseable URL; self-hosted is the default when a mode header is omitted.

For cloud authentication, the HTTP server validates API keys through Parseable Cloud and keeps resolved routing information in a bounded in-memory cache for up to the configured TTL. Self-hosted HTTP requests forward the supplied API key to the selected Parseable instance. Private and loopback target URLs are rejected by default in HTTP mode.

SQL execution is restricted to SELECT queries. DDL and DML are blocked, an automatic LIMIT is applied, and a configurable maximum row count is enforced. Query requests also use a configurable HTTP timeout.

## Setup and configuration

Run `npx -y @parseable/parseable-mcp-server init` for interactive local setup. The command can detect Claude Desktop or Cursor and write client configuration. For scripted stdio setup, provide the client, Parseable URL, and API key explicitly.

Set `PARSEABLE_URL` and `PARSEABLE_API_KEY` for a self-hosted stdio connection. `PARSEABLE_DEFAULT_DATASET` supplies an advisory dataset name, while `PARSEABLE_MAX_ROWS` and `PARSEABLE_QUERY_TIMEOUT_MS` control query bounds and timeout behavior. HTTP mode listens on port 8787 by default and can run from the built Node output or the supplied Docker image.

Cursor's Marketplace plugin uses `PARSEABLE_MODE` and `PARSEABLE_API_KEY`; self-hosted mode also requires `PARSEABLE_URL`. Cloud mode omits the URL. HTTP clients can connect using the `/mcp` endpoint and the documented request headers. Set `PARSEABLE_MCP_ALLOW_PRIVATE=true` only when a trusted deployment must reach private Parseable addresses.

## Limitations and notes

The server does not provide unrestricted SQL execution: write statements and schema-changing statements are rejected. Sampling and query results are bounded by time and row limits. The alert evaluation tool can trigger external notifications. RBAC and cluster tools are read-only, while alert and alert-target tools include configuration changes.

OpenTelemetry export is optional and can send traces to a Parseable OTLP endpoint when enabled. Cloud routing cache entries are disposable, so process restarts or cache misses cause cloud resolution to run again.

_Full upstream README: https://allmcps.com/mcp/parseable-mcp-server/readme_