---
name: pendium-mcp
description: Use when working with the Pendium MCP for AI visibility monitoring, content engineering, or brand analysis. Also use when the user mentions "visibility scan", "AI visibility", "brand score", "content strategy from Pendium data", or "Pendium MCP".
---

# Pendium MCP — Workflow Guide

Pendium monitors how ChatGPT, Claude, Gemini, and AI Overviews perceive and recommend brands. The MCP exposes 18 tools for visibility scanning, content strategy, knowledge queries, and brand analysis. Authentication is handled automatically via OAuth.

## Always Start Here

Call `get_account` first. It returns credit balance, plan tier, and all existing brand agents with their latest scores. This tells you whether you can scan, what agents already exist, and how many credits are left.

## Quick Brand Check (Free)

Use `lookup_brand_score` before paying for a full scan. It searches the public Brand Index by URL, name, or slug — no auth needed. If the brand has been scanned before, you get scores, competitors, and insights instantly.

Only trigger `scan_visibility` when the user needs a fresh, full scan with raw LLM responses.

## Visibility Scan Workflow

1. **Trigger:** `scan_visibility` with URL (new brand) or `syntheticId` (existing agent). Use `mode: "batch"` and `maxQueries: 30` as defaults — good balance of coverage and cost.
2. **Poll:** `get_scan_status` every 30–60 seconds. The `progress` object shows `completedQueries / totalQueries`. Do not call more frequently.
3. **Results:** When `scanAvailable: true`, call `get_report` for the full report with per-query analysis, platform scores, competitor data, and recommendations.
4. **History:** `get_scan_history` shows past scan scores and computed `scoreTrend` (improving/declining/stable).

## Content Engineering Workflow

Use these tools together to write data-grounded, on-brand content:

1. `get_goals_and_ideas` — strategic content goals with specific pitches. Prioritize ideas where `usedCount = 0`.
2. `get_brand_voice` — tone traits, style guide, writing examples, company profile. Load this before writing anything.
3. `get_factsheet` — comprehensive brand overview and key facts for accuracy.
4. `query_knowledge` — semantic search over the brand's knowledge base (website content, uploaded docs) for specific claims and data points.
5. `list_content_briefs` — content plans in the pipeline.

## Action Queue Workflow

1. `get_visibility_actions` — prioritized action items from scan recommendations. Filter by `status` or `priority`.
2. Work through items — create content, do outreach, implement technical fixes.
3. `complete_action` — mark each action done to keep the queue accurate.

## Configuration

- `list_topics` / `add_topic` — view and add visibility topics with search queries.
- `list_personas` / `add_persona` — view and add buyer personas. Personas can be AI-generated from a prompt or defined manually.
- `list_knowledge_sources` — view ingested content sources (website pages, uploaded documents).

## Common Mistakes

| Mistake | Fix |
|---------|-----|
| Scanning without checking credits | Call `get_account` first |
| Using `scan_visibility` for a quick check | Use `lookup_brand_score` (free, instant) |
| Writing content without brand voice | Load `get_brand_voice` before writing |
| Polling too fast | 30–60s for scans |
| Assuming free users can scan via MCP | Free plan can only use `lookup_brand_score`; scans require Starter or Pro plan |