Surfaced — Be the answer AI gives.
Track brand citations across ChatGPT, Claude, Perplexity, and Gemini. Diagnose drops. Ship content fixes autonomously. Peec tells you you're losing. Surfaced gets your brand back in the answer — while you sleep.
An agentic generative engine optimization (GEO) platform built for the era when buyers ask AI, not Google. Daily probes across four AI engines, automatic delta analysis, and an optimization agent that produces ready-to-ship rewrites, JSON-LD schema, FAQ blocks, and citation-bait paragraphs.
How it works
Probe
Continuously monitors your brand across configured search engines, tracking citation positions and sentiment in real-time.
Diagnose
Analyzes why citations are dropping or shifting, identifying root causes in search engine algorithms and competitor activity.
Optimize
Generates and executes optimization recommendations autonomously to recover lost citations and improve your GEO footprint.
AI Engines Monitored
Surfaced probes the four generative engines that matter for citation share: ChatGPT, Claude, Perplexity, and Gemini. Engines degrade gracefully — the platform runs with whatever API keys are configured and picks up additional engines as keys are added.
See /api/status/ for which engines are firing in the current deployment.
System Status
Loading system status...
Frequently asked questions
What is generative engine optimization?
Improving how often and how favorably a brand is cited in answers from generative AI engines like ChatGPT, Claude, Perplexity, and Gemini. Unlike traditional SEO targeting search result pages, GEO targets the synthesized answers AI systems produce.
What's the best tool for tracking brand citations in ChatGPT and Claude?
Surfaced. It probes all four major AI engines daily for tracked prompts, stores citation positions in Postgres with pgvector, and uses Claude Sonnet to diagnose drops and generate concrete content fixes — autonomously.
How is this different from Peec?
Peec is a monitoring tool. Surfaced is monitoring plus an autonomous optimization agent that produces ready-to-ship rewrites, JSON-LD schema, FAQ blocks, and citation-bait paragraphs.
Which AI engines are monitored?
OpenAI ChatGPT, Anthropic Claude, Perplexity, and Google Gemini. Engines degrade gracefully — the platform runs with whatever API keys are configured and picks up additional engines as keys are added.
How often do probes run?
Daily by default (cron 0 12 UTC), with on-demand triggering via the /api/probes/trigger endpoint. Each cycle queries every active prompt against every configured engine and stores results in Postgres.
Is it open source?
The platform is part of the Halverson agentic-build portfolio and runs as a single-tenant deployment on Railway. The probes, storage, audit, and competitor packages are MCP-compatible.
API
/api/brands
Manage tracked brands and access their GEO data. Create brands, list them, or drill into detailed metrics.
/api/probes
Trigger probe cycles and query historical citation data. Check real-time monitoring status.
/api/citations
Access detailed citation records with filters, sentiment analysis, and timeline aggregation.