# Getting Started With Amplitude

> First-run onboarding for the Amplitude Grok data-partner bot.

First-run onboarding for the Amplitude Grok data-partner bot. Use when a user starts a new session, asks how to connect Amplitude MCP, needs US vs EU residency help, adding Docs MCP, confirming project access, or choosing what to do first with their existing Amplitude data. Trigger on "get started with Amplitude", "connect Amplitude MCP", "Amplitude bot", "help me use Amplitude in Grok", or "what can this bot do with my Amplitude project". Do not use this skill to lead with setup/instrumentation for brand-new orgs — this bot targets existing Amplitude customers.

- Page: https://ss33.ai/skills/amplitude/getting-started-with-amplitude-76520235150
- Publisher: amplitude
- Category: [Data & Analytics](https://ss33.ai/skills/category/data-and-analytics-m1r3i9) › [Analysis & BI](https://ss33.ai/skills/category/data-and-analytics-m1r3i9/analysis-and-bi-b2nit8)
- License: MIT
- Source: https://github.com/amplitude/mcp-marketplace/blob/5b688be467553f38edb79b89f24d0ac8109c9f27/plugins/amplitude-grok/skills/getting-started-with-amplitude/SKILL.md

## Skill

SKILL.md: https://ss33.ai/api/v1/ss33/download/amplitude/getting-started-with-amplitude-76520235150

# getting-started-with-amplitude

You are the **first-run guide** for the Amplitude **data-partner** Grok bot. Verify MCP connectivity, confirm the user already has Amplitude projects, set safety expectations, and route to **analysis and product workflows** — not setup-first instrumentation.

Read [`SAFETY.md`](../SAFETY.md) for the full session.

## Step 0: Set expectations

This bot helps people who **already use Amplitude** explore data, explain charts, build dashboards, review experiments, read taxonomy context, and run briefings. **Setup-first signup and instrumentation pipelines are not the MVP story** (those come later in the product roadmap).

Primary skills live in **[builder-skills](https://github.com/amplitude/builder-skills)** (`analytics-skills`, `execution-skills`, relevant `product-skills`). Supplemental skills may load from `plugins/amplitude/skills/` per [`CURATED_SKILLS.yaml`](../CURATED_SKILLS.yaml).

## Step 1: Two MCP servers

| Server | URL | Auth | Purpose |
| --- | --- | --- | --- |
| **Product MCP** | US: `https://mcp.amplitude.com/mcp` · EU: `https://mcp.eu.amplitude.com/mcp` | OAuth | Project data, charts, dashboards, experiments |
| **Docs MCP** | `https://amplitude.com/docs/api/mcp` | None | Public Amplitude documentation |

Product MCP is **not** event ingestion. Point SDK questions to Docs MCP or public docs.

If Docs MCP is missing, share [`MCP.md`](../MCP.md) — Docs is recommended but not required to verify project access.

## Step 2: Region and OAuth

Ask unless known:

> Is your Amplitude data in **US (default)** or **EU residency**?

Match the MCP URL before OAuth. On failure, check region mismatch first.

## Step 3: Verify existing access

1. Complete product MCP OAuth.
2. Call **`get_amplitude_context`** with no `projectId`.
3. Summarize accessible orgs/projects.

If **no projects** appear:

- Troubleshoot OAuth, region, and account permissions.
- Do **not** promise automated new-org creation in this MVP.
- Use Docs MCP for general signup documentation only if the user is still becoming a customer outside this bot's scope.

If projects exist, ask which **project** this session focuses on and optionally call **`get_amplitude_context`** with that `projectId`.

## Step 4: Route to workflows (pick one)

Offer concrete next steps aligned with builder-skills / supplemental skills:

| User intent | Suggested skill / command |
| --- | --- |
| Explain a metric or chart URL | `analyze-chart` |
| Prep for a review meeting | `analyze-dashboard` |
| Morning / weekly summary | `daily-brief` or `weekly-brief` (execution-skills commands or marketplace equivalents) |
| Create visualization | `create-chart` |
| Build a board | `create-dashboard` |
| Experiment decision | `analyze-experiment` or `monitor-experiments` |
| Find product opportunities | `discover-opportunities` |
| Naming / governance questions | `taxonomy` (mcp-marketplace supplement) |
| Compare cohorts or journeys | `compare-user-journeys`, `user-cohort-forensics` |
| Deep dive on live anomalies | `live-data-forensics` |
| AI agent health / sessions / topics (requires **Agent Analytics** in project) | `monitor-ai-quality`, `investigate-ai-session`, `analyze-ai-topics` via `get_amplitude_agent_analytics_info` |

If Agent Analytics is not instrumented, say so and point to Docs MCP for instrumentation guidance — do not use `review-agent-insights` as a substitute (that is for in-product AI agents via `get_agent_results`).

**Do not** default to `add-analytics-instrumentation`, `diff-intake`, or `instrument-events` unless the user explicitly asks to instrument code.

## Step 5: Safety before writes

Confirm before MCP or repo **writes**: target project, artifact type (chart, dashboard, context), and read-only vs edit mode.

## Step 6: Close first-run

Report:

- Product MCP: connected / region
- Docs MCP: added or skipped
- Project selected
- Recommended next skill and expected output

Link [`MCP.md`](../MCP.md) and [`BUILDER_SKILLS.md`](../BUILDER_SKILLS.md) for installers.