Overview
Anycast ActiveEdge API metrics, active session, and live test bench.
Issue sub-10ms Anycast key bound to D1 SQLite FTS5 database.
Quick copy command for Node, Python, or Claude Desktop MCP.
Send live edge query to verify <15ms Anycast response latency.
Crawl documentation or save structured facts for your AI agent.
Live API Test Bench
// Click "Run Test" above to test the endpoint live and inspect edge latency. // Real results from Sovereign Anycast Edge will appear here formatted.
Integration Snippets
# Fastest Web Search API (<15ms)
curl -X POST "https://app.ultrademon.site/v1/search" -H "Authorization: Bearer ud_live_demo_console_key" -H "Content-Type: application/json" -d '{"query": "fastest web search api for ai agents", "limit": 5, "mode": 1}'
from ultrademon import UltraClient
client = UltraClient(api_key="ud_live_demo_console_key")
results = client.search("fastest web search api", limit=5)
print(results)
import { UltraClient } from '@ultrademon/sdk';
const client = new UltraClient({ apiKey: 'ud_live_demo_console_key' });
const searchRes = await client.search('fastest web search api');
console.log(searchRes.results);
{
"mcpServers": {
"ultrademon": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://app.ultrademon.site/mcp", "--header", "Authorization: Bearer ud_live_demo_console_key"]
}
}
}
Summon pre-warmed V8 isolates with zero-cold-start edge key generation.
Live Anycast Latency & Node Diagnostic Probe
Send real live requests to your nearest Cloudflare edge node to measure physical RTT and Ray ID.
| Time | Target | PoP Colo | RTT | Status |
|---|---|---|---|---|
| Click 'Probe Latency' to execute a live Anycast probe. | ||||
Zero-Error Resilience Engine
UltraDemon eliminates API failures through transparent 3-layer edge caching with self-healing circuit breaking.
from ultrademon import UltraDemon
# 0-Cold-Start Anycast Edge Tunnel
demon = UltraDemon.summon(api_key="ud_summon_...")
results = demon.search("autonomous agents", limit=5)
print(results.latency_ms, results.hits)
Interactive Edge Latency Suite
5 Live ProbesFire sequential live HTTP probes from your browser to your nearest Cloudflare edge PoP using your active API key.
Head-to-Head Web Search API Benchmarks
10,000 Real Query Runs AuditedEmpirical comparisons between UltraSearch and centralized search APIs (Tavily, Exa, Serper, Bing Search).
| Engine | Architecture | P50 Latency | P99 Tail | Cold Start | Pricing (100k) | Recall@10 | Sovereignty |
|---|---|---|---|---|---|---|---|
| UltraSearch Edge | Cloudflare Workers + D1 FTS5 | 11.8ms | 38.2ms | 0.0ms | $20.00 | 94.8% | 100% Self-Contained Edge |
| Tavily Search | Centralized AWS us-east-1 | 842.0ms | 2,450.0ms | 320.0ms | $100.00 | 81.4% | Proprietary Cloud |
| Exa (Metaphor) | US West Central Cluster | 910.0ms | 2,800.0ms | 410.0ms | $150.00 | 84.1% | Proprietary Cloud |
| Serper Google API | Scraping Proxy Fleet | 580.0ms | 1,890.0ms | 180.0ms | $100.00 | 79.5% | Proxy Relay |
| Bing Web Search | Azure Central Region | 490.0ms | 1,620.0ms | 250.0ms | $250.00 | 83.0% | Centralized Azure |
Agent Memory Architecture Benchmarks
Sub-15ms TargetComparing UltraMemory against leading AI agent memory solutions (Mem0, Pinecone Serverless, Supabase pgvector).
| Memory Engine | Storage Substrate | Recall Latency (P50) | P99 Tail | Cost / 100k Ops | Local Persistence | Compaction Method |
|---|---|---|---|---|---|---|
| UltraMemory <15ms | D1 SQLite FTS5 + Edge KV | 8.4ms | 24.1ms | $20.00 (Included) | Local + Edge Synchronized | Dynamic BM25 Recency |
| Mem0 | External Vector DB + API | 340.0ms | 1,120.0ms | $80.00 | Central Cloud Only | Embedding Graph |
| Pinecone Serverless | AWS S3 + Pod Partition | 195.0ms | 680.0ms | $70.00 | Cloud Only | Hierarchical HNSW |
| Supabase (pgvector) | Postgres RDS us-east-1 | 135.0ms | 450.0ms | $50.00 | Self-Hosted or Cloud | IVFFlat / HNSW |
Cost Economics & Pricing Model
$0.0002 / ReqUltraDemon utilizes serverless edge economics to deliver 5x to 12x lower costs compared to legacy centralized APIs.
Retrieval Precision & SLA Compliance
BEIR StandardBenchmark results evaluated on standard Information Retrieval benchmarks measuring ranking precision and edge uptime.
Provision Sovereign API Key
Sub-10ms edge validation with native scope for Custom MCP, Seamless Summoning, Search, or Memory.
Active Sovereign API Keys
Pre-warmed across 330+ PoPs| API Key / Prefix | Scope | Plan | Label | Created | Actions |
|---|---|---|---|---|---|
Custom Model Context Protocol (MCP) Hub
Protocol 2024-11-05Connect UltraDemon directly into Claude Desktop, Cursor, Antigravity, and AI agents using native MCP. Run sub-50ms web search and sub-15ms agent memory tools with zero middleware.
// Ready. Select a tool above and click 'Execute MCP Tool'.
UltraMemory API Reference
Store memory, message turns, or batch array into D1 FTS5.
Instant BM25 recall with ranked items and LLM prompt context.
Direct agent prompt context block for zero-overhead LLM injection.
# 1. Store memory from agent
curl -X POST "https://app.ultrademon.site/v1/memory" \
-H "X-API-Key: um_your_key_here" \
-H "Content-Type: application/json" \
-d '{"content": "User prefers Michelangelus font and monochrome UI", "container_tag": "agent_prod"}'
# 2. Recall memory in <15ms
curl "https://app.ultrademon.site/v1/memory/recall?q=user+preferences&container_tag=agent_prod" \
-H "X-API-Key: um_your_key_here"
Store Agent Memory
Persist knowledge, conversation turns, or user preferences into Cloudflare D1 FTS5 with sub-15ms edge persistence.
Recall Agent Memories (BM25 Match)
Instant BM25 Porter-stemmed recall with sub-15ms edge latency across active namespaces.
Stored Memories Explorer
Browse and manage active memory records stored across edge isolate containers.
| Memory ID | Scale | Container | Type | Content | Delete |
|---|---|---|---|---|---|
Mode 1: Instant (<15ms) · Mode 2: Hybrid BM25 (<40ms) · Mode 3: Deep Multi-Pass (<80ms)
UltraSearch API & SDK Quickstart
<40ms BM25 EdgeDirect BM25 web index queries with compiled SQLite FTS5. Zero third-party LLM latency, zero cloud markups.
import { UltraSearch } from '@ultrademon/sdk';
const search = new UltraSearch({ apiKey: 'YOUR_SEARCH_KEY' });
const { results, latency_ms } = await search.query('LLM edge retrieval', {
mode: 1 // 1: Instant <50ms, 2: Standard, 3: Deep
});
import requests
res = requests.get("https://app.ultrademon.site/v1/search",
params={"q": "LLM edge retrieval", "mode": 1},
headers={"X-API-Key": "YOUR_SEARCH_KEY"}).json()
print(f"Results in {res['latency_ms']}ms: {res['results']}")
Test Sovereign Web Index
Execute BM25 keyword matching or deep multi-hop research queries directly against Cloudflare edge nodes.
Ingest URL into Sovereign Index
Autonomous Edge CrawlerIngest URL into Sovereign Index
Fetch, strip boilerplate, extract links, and index live into D1 SQLite FTS5 in real time.
Sovereign Index Explorer
Live IndexLive index corpus stored and searched in Cloudflare D1 across 330+ Anycast edge cities.
| Page Title & Target URL | Domain | Snippet Preview | Words | Crawled | Actions |
|---|---|---|---|---|---|
| Queue Target URL | Domain | Depth & Priority | Status | Added |
|---|---|---|---|---|
API Quotas & Request Credit Packs
UltraDemon operates on a token-style usage model. Top up request credits on demand with zero monthly subscription lock-in, or upgrade to custom business tiers for scaling fleets.
100,000 Requests
- 100,000 unified API requests
- Sub-15ms BM25 edge search
- Persistent SQLite FTS5 memory
- 50 QPS burst capacity
- Zero monthly fee · Top up anytime
Business Scale
- High-volume pools (1M - 100M+ reqs)
- 150+ QPS burst & load balancing
- Dedicated memory namespaces
- Pooled team organization keys
- Autonomous scheduled crawler
Enterprise Sovereign
- Unlimited requests & bespoke economics
- Dedicated Cloudflare PoP sharding
- Private VPC peering & sovereign on-prem
- 99.99% uptime contractual SLA
- Dedicated 24/7 hotline & Architect
Ecosystem SDK Integration Guides
Claude Desktop MCP Setup
{
"mcpServers": {
"ultrademon": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-sse", "https://app.ultrademon.site/mcp/sse"],
"env": { "ULTRASEARCH_API_KEY": "us_your_key_here" }
}
}
}
Interactive Telemetry Simulator & Event Emitter
Instant VerificationTest tracking and micro-metric capture in 1-click. Emits real-time edge events to verify KPIs, kinetic dwell, and VIP trace streams instantly.
UltraMetrics Telemetry API & SDK Quickstart
Sub-5ms Edge IngestDrop our 2.5KB script into any web app to capture kinetic dwell, unmasked GPU silicon, and autonomous customer profiles.
<script src="https://app.ultrademon.site/sdk/ultrademon.js" data-key="YOUR_API_KEY" async></script>
Ideal Customer Profile: Synthesizing...
Clustering recurring visitors, high-affinity traits, and conversion patterns across edge telemetry.
Awaiting visitor telemetry data. Embed the Sovereign Telemetry SDK snippet or emit events to /v1/analytics/capture to automatically synthesize ideal customer traits, circadian habits, and daily hardware abilities.
Daily Abilities & Hardware Arsenal
Daily Habits & Circadian Rhythm
Top Visited Pages
Path / Dwell| Path | Views | Visitors | Avg Dwell |
|---|---|---|---|
| No page views recorded yet. | |||
Traffic Origins & Referrers
Source / Cohort| Referring Domain | Visits | Unique Users |
|---|---|---|
| No referrers recorded yet. | ||
Geographic & Edge PoP Footprint
Anycast PoPs| Location | Edge PoP | Events | Visitors |
|---|---|---|---|
| No geo telemetry recorded yet. | |||
Devices & Platforms
OS / Browser| Platform | Browser | Device | Events |
|---|---|---|---|
| No devices recorded yet. | |||
Insane Micro-Metrics & Kinetic Dwell Telemetry
Sub-Millisecond EdgeCaptures micro-metrics invisible to conventional trackers: active cursor interaction vs idle dwell, unmasked GPU silicon, battery telemetry, and Retina density.
| Activity Classification | Active Dwell | Idle Dwell | Idle Ratio | Events |
|---|---|---|---|---|
| No kinetic activity recorded yet. | ||||
| GPU Silicon / Renderer | Visitors | Adoption |
|---|---|---|
| No GPU telemetry recorded yet. | ||
VIP Client & Specific IP Tracing Engine
Enterprise WatchlistTrack specific high-value IP addresses, target accounts, or individual high-paying clients with custom API keys and real-time micro-trace telemetry.
| Target | Client / Label | Hits Traced | Actions |
|---|---|---|---|
| No watched VIP clients registered. | |||
Integration Snippet (Embed in Any Website or App)
Drop this lightweight 2.5KB script into your HTML <head> or install via npm/python to track visitors and synthesize customer profiles automatically.
<script src="https://app.ultrademon.site/sdk/ultrademon.js" data-key="YOUR_API_KEY" async></script>
// Track custom actions & identify users
window.ultrademon.track('created_project', { plan: 'pro' });
window.ultrademon.identify('usr_9281', { email: 'ceo@startup.com' });
import requests
requests.post("https://app.ultrademon.site/v1/analytics/capture", json={
"api_key": "YOUR_API_KEY",
"event_name": "pipeline_complete",
"distinct_id": "usr_9281"
})
Quickly configure method, headers, and test payloads
HTTP Request Configuration
Edge Response Console
Select a preset above or enter an endpoint and click "Execute Live Request" to test any UltraDemon API live.
Privacy & Cookie Preferences
GDPR · CCPA CompliantUltraDemon uses essential cookies and edge telemetry to maintain session state, guarantee sub-10ms search latency, and sync memory. We never sell your data or use third-party ad tracking.
Sign In to UltraDemon
Enter your email and password to access your workspace
Permanently Delete Key
Target Key: --
This action is irreversible. The key will be permanently expunged from Cloudflare D1 and edge caches across 330+ locations. Enter your account password to confirm deletion.