Loading ToolBench
MCP server for accessing Guzu AI visibility data, tracking brand presence across AI platforms (ChatGPT, Perplexity, Gemini, AI Overview, Grok, Claude), and managing brand analysis workflows.
Server has 5 tools with clear verb-noun naming (get_*, compare_*, ask_*, prepare_*). All tools have descriptions (194 - 280 chars, within baseline 34 - 392 range). Input schemas are visible with typed parameters and defaults. However, output schemas are NOT documented, responses are inferred from code but not formally declared in tool definitions. Parameter descriptions are present but sparse (e.g., 'days' lacks range constraints). No error handling guidance, no recovery hints, no permission declarations. Composition is reasonable (each tool has one responsibility), but missing idempotency guarantees and chaining metadata.
Ask a natural-language question about a brand's AI visibility. Gathers the most relevant visibility data for the brand and returns it so you can answer the question with specific numbers and actionable insight. Use this for open-ended questions like "why is my visibility low?" or "what should I focus on to improve?". Requires a brand that has finished analyzing (see check_analyzing_progress).
Compare the brand's AI visibility against its tracked competitors. Returns the competitor ranking (where the brand sits vs. rivals), overall market position, radar metrics, and platform-specific competitor rankings (who leads on each AI platform). Use this to answer "who beats me in AI answers?" and "where do I rank against competitors?". Requires a brand that has finished analyzing (see check_analyzing_progress).
Get the brand's AI-visibility results broken down by individual prompt (the tracked queries), plus performance grouped by category. Use this to answer "which specific prompts/questions does my brand show up on, and which ones is it missing from?" Each entry shows how the brand performed for that prompt across the AI platforms. Requires a brand that has finished analyzing (see check_analyzing_progress).
Get a brand's AI Visibility Score — how visible the brand is across AI platforms (ChatGPT, Perplexity, Gemini, AI Overview). Returns the AI Visibility Index, total citations and mentions, the brand's market position vs. competitors, and a per-platform breakdown showing which AI platforms cite/mention the brand most. This single tool answers "how visible am I?", "which platform am I strongest on?", and "what's my market position?". Requires a brand that has finished analyzing (see check_analyzing_progress).
Output schemas not documented. Tool definitions show input schemas but no formal return type declarations. LLMs cannot plan downstream calls or extract fields without documented output structure.
Parameter constraints missing. 'days' parameter lacks min/max bounds (should be 1 - 365 or similar). No enum constraints on geography/language. LLMs may pass invalid values.
No error recovery guidance. Tools raise ValueError for missing API key or invalid brand, but descriptions do not explain what to do next or how to recover. No actionable error messages.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | 2026-07-28+ | v2 |
STEP 1 of brand setup. Scan a website and build a draft AI-visibility tracking profile for review. This does NOT start tracking yet — it fetches information about the website so the user can review and edit the offers, competitors, and prompts before committing. It returns: the brand profile (name, type, value propositions), priority offers, competitors, and a set of tracking prompts (the questions Guzu will later ask the AI platforms about this brand).
No permission declarations. Tools do not declare required scopes (e.g., 'read:brand', 'read:visibility'). Audit trail and least-privilege configuration not supported.
Idempotency not guaranteed. Tools do not declare whether repeated calls with same input produce same result. Agents may retry on ambiguous failures, risking duplicate side effects.