Run a live check
POST /api/v1/ai-visibility queries ChatGPT, Claude, Perplexity, Gemini, and Google AI Overview. The response separates completed, partial, and blocked outcomes and reports the brand and domain used for matching.
For Developers
Foglift is built for running agents on AI visibility: its REST API, CLI, and MCP server let Claude Code, Codex, or your own agent check how ChatGPT, Claude, Perplexity, Gemini, and Google AI Overview describe a brand, diagnose why, and fix it. Developer access starts on Launch at $49/mo. The hosted MCP server also works as a ChatGPT connector and a Claude connector.
Best for: running agents on AI visibility.
Free tier requires no API key. Paid plans start at $49/mo.
Foglift exposes the answer evidence behind its dashboard. Pass up to 20 custom prompts to /api/v1/ai-visibility with a selected engine set. Omit the prompts field to run every active saved prompt in the workspace. Each result returns the engine, prompt, whether the brand was mentioned, mention position, sentiment, response text, cited URLs, competitor mentions, and an error field.
POST /api/v1/ai-visibility queries ChatGPT, Claude, Perplexity, Gemini, and Google AI Overview. The response separates completed, partial, and blocked outcomes and reports the brand and domain used for matching.
GET /api/v1/monitor returns 7-, 14-, or 30-day totals, per-engine mention counts, sentiment totals, top citations, and recent results for one brand workspace.
GET /api/v1/history returns paginated answer records with prompt, engine, response text, mention position, sentiment, cited URLs, competitor mentions, and timestamps.
Developer access starts on Launch at $49/mo with 50 API requests per day, 1,000 per month, and daily monitoring across all five engines. Growth raises the limits to 200 per day and 5,000 per month with twice-daily monitoring.
One of our paying customers shipped from signup to first scan in 6 minutes: Stripe checkout, API key, CLI install, working integration. Pick the surface that matches where you already work.
Curl a public Technical Audit. No key needed.
curl --get "https://foglift.io/api/v1/scan" \
--data-urlencode "url=https://yourdomain.com" \
--data-urlencode "format=json"Response shape:
type ScanResponse = {
url: string;
scores: {
overall: number;
seo: number;
aiReadiness: number;
aiVisibility: number;
geo: number;
aeo: number;
performance: number;
security: number;
accessibility: number;
};
topIssues: Array<{
severity: "error" | "warning" | "info";
category: string;
title: string;
}>;
totalIssues: number;
scanId: string;
fullReport: string;
};Run from terminal or CI. Colored output, JSON, threshold gating.
npm install -g foglift-scan
foglift scan https://yourdomain.com --json30 hosted MCP tools available inside your AI editor.
Add Foglift to ChatGPT or Claude as a connector
Read the Google Search Console MCP server guide
// claude_desktop_config.json
{
"mcpServers": {
"foglift": {
"command": "npx",
"args": ["foglift-mcp"]
}
}
}Record AI crawler hits from your own site.
npm install @foglift/sensor@^1.0.0
import { senseAITraffic } from "@foglift/sensor/nextjs";
export const middleware = senseAITraffic({
siteToken: "YOUR_WORKSPACE_ID",
});Three patterns we see most often from teams that wire Foglift in directly.
Scan every pull request against your tracked prompts. Fail the build if AI Readiness drops below threshold. Same pattern as Lighthouse CI, applied to AI search.
# .github/workflows/foglift.yml
- run: npx foglift-scan scan https://preview-${{ github.sha }}.example.com --threshold 75Pull visibility history, sentiment, and citation data from /api/v1/historyinto Grafana, Metabase, your internal BI tool, or a Notion database. AI search visibility becomes a metric your team already watches alongside latency and conversion.
curl -H "X-API-Key: $FOGLIFT_KEY" \
"https://foglift.io/api/v1/history?brand_id=$BRAND_ID&limit=100" \
| jq '.results[] | {date, visibility_rate, by_model}'Replace $BRAND_ID with your brand UUID from /dashboard/developer. Required on multi-brand accounts; single-brand accounts auto-resolve.
Give Claude Code, Codex, Cursor, or your own agent one bounded six-step loop:
# In Cursor or Claude Code, with the MCP server installed:
"Scan /pricing on the preview deploy, then add FAQ schema to fix the
top recommendation, then re-scan and report the delta."Nine endpoints cover the full platform. Full reference in /docs.
| Endpoint | Verb | Purpose |
|---|---|---|
| /api/v1/scan | GET | Public technical audit. No key. AI Readiness + SEO + security + a11y scores in JSON or text. |
| /api/v1/scan/batch | POST | Scan up to 10 URLs in one request. API key required. |
| /api/v1/aeo-score | GET | Detailed AI Readiness breakdown across 8 dimensions, grade, top recommendations. |
| /api/v1/content-brief | GET | Target prompts, content gaps, structure recommendations for a page. |
| /api/v1/monitor | GET | Read AI Visibility results, model breakdown, citations, and trends. Launch+. |
| /api/v1/ai-visibility | POST | Query ChatGPT, Perplexity, Claude, Gemini, and Google AI Overview for up to 20 custom prompts, or omit prompts to run every active saved workspace prompt. Debits tokens per model. Launch+. |
| /api/v1/prompts | GET/POST/DELETE | CRUD for monitoring prompts. API key required. |
| /api/v1/history | GET | Paginated visibility history with per-model breakdown. |
| /api/v1/results/{result_id}/trace | GET | Read ordered ChatGPT search steps, pages read, citation flags, and the 30-day prompt rollup. Launch+. |
| /api/v1/crawler-analytics | GET | AI bot visits to your site (GPTBot, ClaudeBot, PerplexityBot, etc.). |
Authenticate with an X-API-Key header. Generate keys and copy brand UUIDs in Dashboard → Developer.
The Launch plan exists for solo founders and small teams who want the full API, MCP, and CLI surface without committing to an agency-tier price.
Launch
API + MCP + CLI
Growth
Twice-daily monitoring
Enterprise
Read + write API
Free tier (no API key) gets public scan access with per-IP limits. Full plan comparison on /pricing.
Short answers for agents and builders deciding whether Foglift has the integration surface they need.
Yes. Foglift exposes a public REST API for Technical Audits, AI Visibility monitoring, prompt management, history, crawler analytics, and related account data. The public /api/v1/scan endpoint works without an API key; authenticated endpoints use an X-API-Key header generated in Dashboard -> Developer.
API-key access is available on Launch, Growth, and Enterprise plans starting with Launch at $49/mo. The public Technical Audit endpoint remains available with per-IP limits, while Launch+ adds higher limits, multi-engine AI Visibility monitoring, CLI workflows, and the hosted MCP server.
POST /api/v1/ai-visibility accepts a custom prompts array with up to 20 entries. Omit the prompts field to run every active saved prompt in the workspace. Each result returns the engine, prompt, whether the brand was mentioned, mention position, sentiment, response text, cited URLs, competitor mentions, and an error field. GET /api/v1/monitor adds 7-, 14-, or 30-day totals, per-engine mention counts, sentiment totals, and top citations.
Yes. The published npm package is foglift-scan, and it installs the foglift binary for terminal and CI workflows. Run npm install -g foglift-scan, then foglift scan https://yourdomain.com to get JSON or text Technical Audit output.
Yes. Foglift ships a Model Context Protocol server through the foglift-mcp npm package and the hosted https://foglift.io/api/mcp endpoint. The hosted server works as a ChatGPT connector (Plus, Pro, Business, or Enterprise, with Developer mode on) and as a Claude connector on paid Claude plans. Agent clients such as Claude Code, Codex, Cursor, and Windsurf can also call Foglift tools for scans, AI Visibility checks, recommendations, referrer analytics, recipes, and setup guidance.
Generate an API key in your dashboard, open the API docs, or read how AI agents discover executable MCP and API surfaces.
A working series on AI search visibility from the operator’s seat. Written for founders shipping AI-native products who care about how their own brand surfaces in ChatGPT, Claude, Perplexity, Gemini, and Google AI Overview.
Live
Compare APIs, CLIs, MCP servers, and CI-friendly workflows for testing and improving AI search visibility.
Live
See which brands ChatGPT recommends, which sources support them, and how to turn the missing evidence into a measurable improvement plan.
Live
Week-by-week plan covering setup, baseline measurement, content shipping, citation building, and the metric you measure on day 90.