API Reference

Glance enriches public URLs and returns an AI-improved title, description, content type, a Glance-generated OG preview image, and optional safety and injection metadata.

Two enrichment modes: /v1/enrich (Basic — fast fixed pipeline, all plans) and /v1/intelligence (agent with trust scoring and audit trail, Growth+). Both reject private-network and metadata URLs before any fetch.

Try the live playground without an API key, or sign in to get a key and use the authenticated endpoints.

Base URLs

Use the same paths on either host. This site also rewrites /v1/* to /api/v1/* so both work on useglance.io.

Website + app-hosted API

https://www.useglance.io/api/v1

Shorthand: https://www.useglance.io/v1 (same routes). /v1/intelligence is available on this host only.

Production API (Cloudflare Worker)

https://api.useglance.io/v1

Persistent keys, batch jobs, and OG images on cdn.useglance.io when the worker is deployed. Basic enrich and safety only — no /v1/intelligence on this host yet.

Authentication

Authenticated endpoints require a Bearer token. The playground is the only public enrich path.

Authorization: Bearer glance_your_api_key_here

Sign in at /login with your email and API key. Keys are shown and copied from the dashboard. Key rotation is not available in the dashboard yet.

Playground

Try enrichment without auth. Does not count toward monthly quota. Same URL rules as enrich — only public http and https URLs; private IPs and localhost are rejected.

GET/v1/playground

Parameters

Param	Type	Description
`url`	string	Required. URL-encoded public http(s) URL. Private networks and metadata endpoints are blocked.

Example

curl "https://useglance.io/v1/playground?url=https%3A%2F%2Fgithub.com%2Fvercel%2Fnext.js"

Enrich a URL

GET/v1/enrich

Scrapes the page, classifies content, improves title and description with a cloud LLM, and renders a 1200×630 PNG card using your template settings. Always returns a Glance-generated og_image (not the site's original OG file). Private IPs, localhost, and cloud metadata URLs are rejected before any fetch. Page fetches run in an isolated Browserless sandbox in production. Basic pipeline: scrape → classify → OG card. Also scans page text for prompt-injection patterns on every request (injection_risk). Counts toward monthly quota on all plans.

Parameters

Param	Type	Description
`url`	string	Required. URL-encoded public http(s) URL. Private IPs, localhost, and cloud metadata URLs return 400.
`safety`	boolean	Optional. Pass "true" to include safety fields on the enrich response.

Example

curl "https://useglance.io/v1/enrich?url=https%3A%2F%2Fgithub.com%2Fvercel%2Fnext.js&safety=true" \
  -H "Authorization: Bearer glance_..."

URL Intelligence (Growth+)

Agent-powered enrichment for AI apps and security-sensitive workflows. Instead of a fixed pipeline, Glance runs a reasoning agent that gathers signals — Google Safe Browsing, VirusTotal, prompt-injection scan, domain age, archive fallback for paywalls — then classifies content and returns an auditable intelligence block with trust score and step-by-step trace.

Use when you need to safely preview links from user input, RAG citations, or agent tool calls. Basic /v1/enrich remains available on all plans for fast metadata. Requires Growth or Pro — Free and Starter receive 403 plan_required.

Host: https://www.useglance.io/v1/intelligence (app deployment only).

GET/v1/intelligence

Parameters

Param	Type	Description
`url`	string	Required. URL-encoded public http(s) URL. Same SSRF rules as enrich.

Example

curl "https://www.useglance.io/v1/intelligence?url=https%3A%2F%2Fexample.com" \
  -H "Authorization: Bearer glance_..."

Errors

If the agent flags a URL as dangerous, the response is 403 unsafe_url with an optional intelligence.steps array showing how far the agent got before blocking.

Response (excerpt)

{
  "url": "https://example.com",
  "title": "...",
  "confidence": 0.82,
  "mode": "intelligence",
  "intelligence": {
    "trust_score": 0.78,
    "injection_risk": "none",
    "domain_age_days": 9125,
    "confidence_reasoning": "High confidence — safety, domain, and content signals align.",
    "used_cached_content": false,
    "slug_matches_content": true,
    "steps": [
      { "tool": "fetch_page", "status": "ok", "summary": "Fetched live page", "duration_ms": 420 },
      { "tool": "check_safety", "status": "ok", "summary": "Risk: safe (0.02)", "duration_ms": 180 },
      { "tool": "scan_injection", "status": "ok", "summary": "No injection patterns detected", "duration_ms": 2 }
    ]
  }
}

Safety check

Standalone risk assessment via Google Safe Browsing, VirusTotal (70+ engines), and heuristics. Counts as one enrichment toward your quota. Only public http(s) URLs accepted.

Uses VirusTotal url-info (GET cached report — fast, 1 lookup). Optional scan-url fallback when VIRUSTOTAL_SCAN_ON_MISS=true (submits unknown URLs, uses extra quota). Cached 6h, max 4 live calls/min on free tier.

GET/v1/safety

Example

curl "https://useglance.io/v1/safety?url=https%3A%2F%2Fexample.com" \
  -H "Authorization: Bearer glance_..."

Response

{
  "url": "https://example.com",
  "risk_level": "safe",
  "risk_score": 0.02,
  "threat_types": [],
  "signals": [
    { "check": "google_safe_browsing", "result": "clean", "weight": "critical", "passed": true },
    { "check": "virustotal", "result": "clean (68/72 harmless)", "weight": "critical", "passed": true }
  ],
  "cached": false,
  "processing_time_ms": 210
}

risk_level values

`safe`	Score below 0.2, no critical threats.
`suspicious`	Heuristic flags; score 0.2–0.5.
`dangerous`	Safe Browsing match or score above 0.5.

Check usage

GET/v1/usage

curl "https://useglance.io/v1/usage" \
  -H "Authorization: Bearer glance_..."

Response

{
  "user_id": "user_abc",
  "plan": "free",
  "period": "2026-05",
  "count": 42,
  "limit": 500,
  "remaining": 458
}

OG templates

Configure which card layout is used per content_type and (on Growth / Pro) brand colors, logo, fonts, and layout. Saved per API key. Edit in the dashboard or via API.

Built-in templates: article, product, repo, profile, generic. If you do not set a default for a content type, Glance picks one automatically (e.g. repo for GitHub URLs).

Get config

GET/v1/templates

Update config

PATCH/v1/templates

Send a partial JSON body with defaults, theme, and/or layout. Brand fields require Growth or Pro. Glance logo on OG images is always on except Pro (showBadge: false).

curl -X PATCH "https://useglance.io/v1/templates" \
  -H "Authorization: Bearer glance_..." \
  -H "Content-Type: application/json" \
  -d '{
    "defaults": { "article": "article", "repo": "repo" },
    "theme": { "accentColor": "#6366f1", "showBadge": true, "fontFamily": "inter" },
    "layout": { "backgroundStyle": "dark", "titleAlign": "left" }
  }'

Preview

POST/v1/templates/preview

Returns a PNG (not JSON). Same auth as other routes.

Enrich response

{
  "url": "https://github.com/vercel/next.js",
  "title": "Next.js: The React Framework",
  "description": "Production-ready React framework with SSR and routing.",
  "og_image": "https://www.useglance.io/api/v1/img/5973e384f5c0baa1",
  "share_url": "https://www.useglance.io/link?url=https%3A%2F%2Fgithub.com%2Fvercel%2Fnext.js",
  "favicon": "https://github.com/favicon.ico",
  "site_name": "GitHub",
  "content_type": "repo",
  "source": "hybrid",
  "confidence": 0.95,
  "injection_risk": "none",
  "safety": {
    "risk_level": "safe",
    "risk_score": 0.02,
    "threat_types": []
  },
  "cached": false,
  "cached_at": null,
  "processing_time_ms": 1240
}

og_image on the worker deployment uses https://cdn.useglance.io/og/.... share_url is set when the request includes a public site origin (typical on useglance.io). safety is present when safety=true is passed on enrich.

injection_risk

Included on every enrich and intelligence response. Scans page text for prompt-injection patterns before the LLM runs.

`none`	No suspicious patterns detected.
`low`	Minor patterns (e.g. role-override phrasing); confidence may be reduced.
`high`	Strong injection signals; see optional injection.patterns on enrich.

When risk is not none, enrich may also include:

"injection": {
  "detected": true,
  "risk": "high",
  "patterns": ["ignore-instructions", "zero-width-chars"]
}

Intelligence-only fields

/v1/intelligence responses add mode: "intelligence" and an intelligence object with trust_score, confidence_reasoning, domain_age_days, used_cached_content, slug_matches_content, and steps[] (full agent audit trail). See the Intelligence section for an example.

content_type

articleproductrepoprofilelandingvideoother

source

`generated`	Title and/or description mainly from AI and page content.
`existing`	Strong existing OG/meta tags; little or no AI rewrite.
`hybrid`	OG tags present but improved or merged with AI output.

Caching

Results are cached by normalized URL and your template config. Cache hits return cached: true and are much faster.

On this deployment, typical TTLs are shorter (about 6–24 hours depending on content type). The worker deployment uses longer TTLs (up to 30 days for repos and profiles).

Changing template or brand settings bumps the config version so new enrichments get fresh images.

Errors

JSON error bodies include error and message.

HTTP	Code	Meaning
400	`invalid_url`	Missing url, malformed URL, or blocked target (private IP, localhost, metadata endpoint, non-http(s)).
401	`unauthorized`	Missing Bearer token.
403	`invalid_key`	API key invalid or revoked.
403	`plan_required`	Endpoint requires Growth or Pro (e.g. /v1/intelligence on Free/Starter).
403	`unsafe_url`	URL flagged dangerous by safety checks. Intelligence may include steps[] in the body.
429	`rate_limited`	Monthly quota exceeded.
500	`internal_error`	Unexpected failure; retry.

If scraping fails, enrich still returns a minimal result (domain + favicon) when possible rather than an empty 500.

Rate limits

Quotas are per API key per calendar month. There is no per-second rate limit on current plans.

Each GET /v1/enrich, GET /v1/intelligence, and GET /v1/safety counts as one enrichment toward your quota, including cache hits.
GET /v1/playground does not count.
Template GET/PATCH and preview POST do not count toward enrich quota.

See pricing for plan limits.

API Reference

Base URLs

Authentication

Playground

Parameters

Example

Enrich a URL

Parameters

Example

URL Intelligence (Growth+)

Parameters

Example

Errors

Response (excerpt)

Safety check

Example

Response

risk_level values

Check usage

Response

OG templates

Get config

Update config

Preview

Share previews

Enrich response

injection_risk

Intelligence-only fields

content_type

source

Caching

Errors

Rate limits