aigclink/geolook

What it solves

GeoLook is a self-hosted platform designed for Generative Engine Optimization (GEO). It helps brands ensure they are proactively mentioned and cited by AI engines like ChatGPT, Perplexity, Gemini, and DeepSeek when users ask category-related questions. It solves the problem of AI engines acting as "black boxes" by quantifying mention rates, diagnosing why a brand is absent or outranked by competitors, and providing a structured path from diagnosis to execution and verification.

How it works

The platform operates through a four-stage workflow:

1. Status Analysis: It samples answers from 17 different AI engines (via API or a manual Chrome extension) to measure mention rates, citation shares, and brand distribution across different user intents (e.g., buyer, educate, probe).
2. Diagnosis: It performs a site audit based on a four-layer dependency chain (Access $\rightarrow$ Orientation $\rightarrow$ Understanding $\rightarrow$ Quotability). It checks for crawler accessibility, WAF/CDN blocks, and the presence of quotable passages.
3. Action: It generates structured implementation tickets with risk grades and acceptance criteria. It includes a content workbench for drafting AI-friendly content and tools to deploy assets like llms.txt and JSON-LD.
4. Verification: It programmatically verifies fixes through re-crawling and subsequent sampling rounds to measure the actual impact on AI mentions.

Who it’s for

• Marketing agencies and consultants who deliver GEO as a service and need professional client delivery packages (reports, strategy, and execution plans).
• Brand owners who want to increase their visibility in AI-generated answers without relying on a cloud-based SaaS.
• SEO professionals transitioning into AI-driven search optimization.

Highlights

• End-to-End Implementation: Moves beyond simple monitoring to provide a full loop of diagnosis $\rightarrow$ tickets $\rightarrow$ assets $\rightarrow$ verification.
• First-class Chinese Market Support: Includes a comprehensive matrix of CN engines (e.g., GLM, Doubao, DeepSeek) and calibrated citation channels.
• Self-Hosted & Private: All data is stored locally in JSON/Markdown files; no external database or cloud account is required.
• Sampling Assistant: A Chrome extension and sandbox environment that allows for high-quality manual sampling from engines that lack APIs.
• Evidence-Based Scoring: Audit dimensions are based on empirical data regarding word counts, content patterns, and citation probabilities.