Optimize · MCP Server
The Spyglasses MCP server: query your AI visibility data from any AI assistant.
The Spyglasses MCP server connects your account to Claude and any MCP-compatible client. Ask for metrics, citations, goals, and scores in plain language; run AIPVS and PQS scoring; drive the Citation Optimizer loop; and pull raw rows into notebooks, BI tools, and custom workflows. Access follows your Spyglasses login and workspace permissions.
The MCP server at a glance
- Works with Claude and any MCP-compatible client.
- Access follows your Spyglasses login; write-access features will follow the MCP permission policies, so you approve each write.
- Metrics, citations, goals, AIPVS/PQS scoring, and the Citation Optimizer loop as tools.
- Guided prompts cover the common analyses: projects, message drift, publishers.
Queried the Spyglasses MCP server
Top sources: OutdoorGearLab (24), Switchback Travel (17)… AIPVS-weighted, review platforms drove 61% of attributed citations. Spyglasses
19 citations · prompt, platform, date, answer context → guides-goal-citations.csv Spyglasses
What you can query
Your whole account, as tools
Reports, account metrics, citations and goals, publisher and placement scoring, and the score-revise-rescore loop of the Citation Optimizer: each is a tool your assistant can call, and every series comes back as data it can chart.
Reports
Visibility reports and site audits, by share token.
Metrics & citations
Trends, consistency, message tracking, goals, and raw citation rows.
Scoring
AIPVS for publishers and PQS for placements, on demand.
Optimizer loop
Score content, request a revision, re-score, repeat.
Guided prompts
The common analyses, one slash away
Analyze a project, track message drift, evaluate publishers, review citation mix: guided prompts orchestrate the right tool calls so you don't have to know the API to get the analysis.
Build your own views
Raw rows for the analysts
Everything the dashboards show is queryable: pull raw citation rows into a notebook, feed a BI dashboard, or wire Spyglasses data into custom client workflows.
Your data, your rules
Your login, your permissions.
The MCP server uses your Spyglasses login, and access follows your workspace permissions. Today the account surfaces focus on reading your data; as write-access features are introduced they'll follow the MCP permission policies, so you approve each write before it happens.
Agencies: connect client data to your own stack and let analysts self-serve.
For agencies →Connect your assistant.
Setup is an OAuth sign-in away; the docs walk you through it.
MCP server FAQ
Common questions about querying Spyglasses over MCP.
Which assistants work with it?
Claude (claude.ai and Claude Code) and any client that speaks MCP. If your tool supports MCP servers, it can connect.
Is my data safe?
You sign in with OAuth using your own Spyglasses account, and account data is exposed read-only. Access follows your workspace permissions.
Can it change anything in my account?
Not without you approving it. Today the account surfaces focus on reading your data; as write-access features are introduced they'll follow the MCP permission policies, so you approve each write before it happens.
What's the difference from an API?
You don't write code against it; your assistant calls the tools for you. Ask a question in plain language and get charts, tables, or files back.
Can clients get their own access?
Anyone with a Spyglasses login can connect their own assistant; access follows workspace membership, so clients see what you've shared with them.