Connect Claude, Cursor, or VS Code to your Vizzybl workspace with the remote MCP server — one-click OAuth or an API key.
Last updated
Vizzybl runs a remote MCP (Model Context Protocol) server that exposes your workspace's GEO data — and a handful of metered actions — as tools any MCP client can call. Connect it once and your AI assistant can pull share-of-voice trends, read verbatim AI answers, run audits, and scrape prompts without you writing a single API call.
https://mcp.vizzybl.ai/v1/mcpYou need:
OAuth (1-click). Add Vizzybl as a connector in your client and a browser window opens to sign in and pick the workspace to connect. Access tokens last 8 hours and refresh automatically, so you authorize once and forget about it. Behind the scenes, consent mints an API key named MCP: <client> in Settings → API Keys — revoke that key there to disconnect the client at any time.
API key (PAT). Create a key under Settings → API Keys (see API Authentication) and configure your client to send it as Authorization: Bearer vzbl_live_.... Use this for clients without OAuth support, or for headless / scripted setups.
Claude Desktop or claude.ai. Go to Settings → Connectors → Add custom connector, paste https://mcp.vizzybl.ai/v1/mcp, and complete the OAuth consent in the browser.
Claude Code. One command, then approve the OAuth prompt:
claude mcp add --transport http vizzybl https://mcp.vizzybl.ai/v1/mcp
Cursor. Add the server to ~/.cursor/mcp.json:
{"mcpServers":{"vizzybl":{"url":"https://mcp.vizzybl.ai/v1/mcp"}}}
VS Code. Add the server to .vscode/mcp.json in your project:
{"servers":{"vizzybl":{"type":"http","url":"https://mcp.vizzybl.ai/v1/mcp"}}}
Then ask a question in plain language — "How is my brand doing in AI search this month?", "Which competitor domains get cited where we don't?", "Run a GEO audit on our pricing page."
| Tool | What it returns |
|---|---|
get_brand | Brand identity: name, primary domain, variants, competitors, industries, markets |
list_engines | The AI answer engines Vizzybl tracks |
get_quota | Plan tier, credit balance, and per-feature quota usage |
get_visibility_stats | One-call rollup: latest share-of-voice, rank, sentiment, trend, top competitors |
get_visibility_history | Daily share-of-voice history over time |
list_prompts | Monitored prompts with aggregated metrics |
get_prompt | A single prompt with its full metrics |
get_prompt_metrics_history | Daily metrics time-series for one prompt |
get_prompt_responses | Verbatim AI answers with citations, brand mentions, sentiment (untrusted content) |
list_citations | Domains AI engines cite, flagged as yours or a competitor's (untrusted content) |
find_prompts_citing | Reverse lookup — which prompts cite a given URL or domain |
get_citation_stats | Aggregated citations — domain rank and share, most-cited URLs, per-competitor breakdown, source types |
get_recommendations | The recommendation feed — Actions Board items plus nightly AI Findings, priority-sorted (untrusted content) |
get_bot_traffic | AI-crawler traffic from your access logs, grouped by bot, status, path, and day (untrusted content) |
list_tags | Custom tag dimensions and values used to group prompts |
list_audits | GEO audit history for the workspace, newest first |
list_products | Product catalog with each product's latest AI-answer visibility (untrusted rival names) |
get_product_metrics_history | One product's daily visibility series (untrusted rival names) |
| Tool | What it does |
|---|---|
weekly_pulse | Share-of-voice, rank, and sentiment for the last 7 days vs the prior 7 — week-over-week movement at a glance |
citation_gap | Domains AI engines cite where your own domain is not cited — the content-opportunity list |
engine_scorecard | Per-engine visibility: share-of-voice, trend, and mentions for each AI engine over a window |
source_authority | How well your own pages earn citations — own-domain share, most-cited URLs, and source-type mix |
product_scorecard | Which products win or lose in AI answers, and which rival products engines name instead |
| Tool | What it does |
|---|---|
run_audit | Trigger a GEO audit of a URL (async; consumes credits, settled by the audit pipeline) |
get_audit | Poll an audit's status and scores — read-only, no credits |
create_prompts | Add monitored prompts, up to 50 at a time (free; supports a preview dry-run) |
scrape_prompt | Trigger a fresh scrape of a prompt across AI engines (async; charges prompt_monitor_run) |
Write tools are metered — they consume credits from your workspace balance — and MCP clients surface a confirmation prompt before running them, so nothing is spent without your approval.
run_audit and scrape_prompt are metered — they consume credits. create_prompts is free, and get_audit is read-only. MCP clients surface a confirmation prompt before any tool that is not read-only runs.
Tools inherit the scopes of the key behind the connection. Read tools need the matching read: scope (or read:*) — read:brand, read:visibility, read:prompts, read:responses, read:citations, read:recommendations, read:traffic, or read:products. run_audit needs write:audits, and create_prompts / scrape_prompt need write:prompts. A tool called without its scope returns a 403 error naming the missing scope. Use get_quota to check your credit balance before large write actions.
MCP: <client>). Revoke it there to cut off a client immediately.