nexusV1 · production api · 500+ users
lostingness
AI Gateway
Advanced AI API with web search, 94+ models, image generation, vision, and persistent memory — built by @lostingness
POST https://osint.lostingness.site/nexusai.php
✓ 500 Users ✓ Rate Limited ✓ Tavily Search ✓ Serper Google ✓ 94+ Models ✓ Image Gen ✓ Vision/OCR ✓ AI Memory
94+
Text Models
5
Image Models
16
Vision Models
13
API Actions
2
Search Engines
$25
Free/Week
⚡ How nexusV1 Works
Auto-routing, dual search engines, persistent memory — one API call does everything
POST Requesthttps://osint.lostingness.site/nexusai.php │ ▼ Rate Limit (20/min · 200/hr · 800/day per IP) │ ▼ action = "nexusv1"?Auto-detect query type │ ├─ SEARCHgpt-4.5-turbo ├─ CODINGgemini-2.5-flash ├─ REASONINGdeepseek-r1 ├─ CREATIVEgrok-3 ├─ LONGkimi-k2.5 (128k) ├─ VISIONgemini-3-flash-preview └─ GENERALgemini-2.5-flash │ ▼ needs_search()? ← today/aaj/price/news/khabar/2025... YES → Tavily (advanced) + Serper (Google) → merge context NO → skip │ ▼ Load memory.json → inject relevant past Q&A │ ▼ AI call (date + search results + memory + system prompt) │ ▼ Auto-save Q&A → memory.json │ ▼ Clean JSON Response ✅ (APIs masked — only lostingness AI shown)
🧬 nexusV1 Auto-Routing
Paste nexusV1 as the model — it picks the best AI for your task automatically
🔍
SEARCH
nexusV1-search
Real-time news, prices, current events
today · aaj · price · news · khabar · live · 2025
💻
CODING
nexusV1-code
PHP, Python, JS, debugging, dev
code · function · bug · debug · php · python
🧮
REASONING
nexusV1-reason
Math, logic, step-by-step analysis
why · calculate · prove · compare · analyze
CREATIVE
nexusV1-creative
Stories, poems, scripts, shayari
write a story · poem · shayari · kahani
📄
LONG CONTEXT
nexusV1-long
Documents, summaries, deep dives
summarize · full analysis · comprehensive
👁️
VISION
nexusV1-vision
Image analysis, OCR, visual understanding
+ image_url in request
🚀 Quick Start
Working code in 4 languages — copy and run
curl -X POST https://osint.lostingness.site/nexusai.php \
  -H "Content-Type: application/json" \
  -d '{
    "action": "nexusv1",
    "messages": [
      {"role": "user", "content": "Aaj ka gold price kya hai?"}
    ]
  }'
$ch = curl_init('https://osint.lostingness.site/nexusai.php');
curl_setopt_array($ch, [
    CURLOPT_RETURNTRANSFER => true,
    CURLOPT_POST           => true,
    CURLOPT_POSTFIELDS     => json_encode([
        'action'   => 'nexusv1',
        'messages' => [['role' => 'user', 'content' => 'Hello!']],
    ]),
    CURLOPT_HTTPHEADER     => ['Content-Type: application/json'],
    CURLOPT_TIMEOUT        => 60,
    CURLOPT_SSL_VERIFYPEER => false,
]);
$res  = curl_exec($ch);
curl_close($ch);
$data = json_decode($res, true);
echo $data['content'];
const res = await fetch('https://osint.lostingness.site/nexusai.php', {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({
    action: 'nexusv1',
    messages: [{role: 'user', content: 'What is AI?'}]
  })
});
const data = await res.json();
console.log(data.content);
import requests

res = requests.post('https://osint.lostingness.site/nexusai.php', json={
    'action': 'nexusv1',
    'messages': [{'role': 'user', 'content': 'Explain Python asyncio'}]
}, timeout=60)

data = res.json()
print(data['content'])
📦 Response Format
// Success
{
  "ok": true,
  "content": "AI response here...",     // nexusv1 / chat
  "model": "nexusV1",
  "type": "nexusV1-search",             // which sub-model was used
  "query_type": "SEARCH",              // SEARCH/CODING/VISION/REASONING/CREATIVE/LONG/GENERAL
  "search_used": true,                 // web search triggered?
  "sources": ["https://..."],          // search sources
  "has_memory": true,                  // memory context injected?
  "usage": {
    "prompt_tokens": 150,
    "completion_tokens": 80,
    "total_tokens": 230
  },
  "powered_by": "lostingness AI v6.0.0"
}

// Error
{ "ok": false, "error": "message", "code": 400 }

// Rate limited (HTTP 429)
{ "ok": false, "error": "Rate limit message", "rate_limited": true, "retry_after": 60 }
🤖 All Available Models
94+ text models, 5 image models, 16 vision models — all accessible via single endpoint
nexusV1 Custom Models
9 models
RECOMMENDED
Model IDBest ForNotes
nexusV1Everything — auto routing★ START HERE
nexusV1-searchNews, prices, real-time+ Tavily + Serper
nexusV1-codePHP, Python, JS, debugCode expert
nexusV1-visionImage analysis, OCRVISION
nexusV1-reasonMath, logic, deep thinkingREASON
nexusV1-creativeStories, poems, shayariCreative mode
nexusV1-longLong documents, summaries128k context
nexusV1-fastQuick short answersLightweight
nexusV1-imageImage generationzimage engine
Google — Gemini
6 models
Model IDContextBest ForTag
gemini-2.5-flash1MGeneral, code, fast★ BEST CHEAP
gemini-2.5-pro1MComplex, detailed tasksPremium
gemini-3-flash-preview1MVision + generalVISION ★
gemini-1.5-flash1MFast generalStable
gemini-1.5-pro2MHuge context2M tokens
gemma-3128kLightweight openOpen source
OpenAI — GPT / o-series
14 models
Model IDContextBest ForTag
gpt-4.5-turbo128kSearch + reasoning★ SEARCH
gpt-4o128kVision + generalVISION
gpt-4o-mini128kCheap + capable★ CHEAP
gpt-48kPremium qualityClassic
gpt-4-turbo128kFast GPT-4Turbo
gpt-4.11MLatest 1M context1M tokens
gpt-4.1-mini1MFast + cheapMini
gpt-4.1-nano1MUltra fastNano
o1200kDeep step-by-stepREASON
o1-mini128kFast reasoningREASON
o3200kAdvanced reasoningREASON
o3-mini200kFast o3REASON
o4-mini200kLatest miniREASON
o4-mini-high200kHigh effortThorough
xAI — Grok
5 models
Model IDContextBest ForTag
grok-3128kCreative, witty★ CREATIVE
grok-3-mini128kFast creativeLightweight
grok-2-1212128kStable versionDec 2024
grok-beta128kLatest featuresBeta
grok-3-vision-beta128kVision + creativeVISION
Meta — Llama
8 models
Model IDContextBest ForTag
llama-3.1-8b-instruct128kFast lightweight8B
llama-3.1-70b-instruct128kBalanced70B
llama-3.1-405b-instruct128kPowerful Llama405B
llama-3.2-11b-vision-instruct128kVision smallVISION
llama-3.2-90b-vision-instruct128kVision largeVISION
llama-3.3-70b-instruct128kLatest 70BImproved
llama-4-scout10M10M context!Massive
llama-4-maverick1MLatest multimodalVISION
DeepSeek
4 models
Model IDContextBest ForTag
deepseek-v3.2128kCode, analysis★ CODE VALUE
deepseek-r164kMath, reasoning★ REASON
deepseek-r1-distill-qwen-32b32kFast reasoningDistilled
deepseek-r1-distill-llama-70b64kLlama-basedDistilled
Anthropic — Claude
3 models
Model IDContextBest ForTag
claude-3-5-sonnet-20241022200kWriting, analysis, visionVISION
claude-3-haiku-20240307200kFast lightweightCheap
claude-sonnet-4-5200kLatest SonnetNewest
Alibaba — Qwen
6 models
Model IDContextBest ForTag
qwen3-235b-a22b128kGiant 235B modelMassive
qwen3-30b-a3b128kMid-size capable30B
qwen3-8b128kFast lightweight8B
qwen-turbo128kFast + cheapTurbo
qwen-max32kMost capablePremium
qwq-32b32kMath + reasoningREASON
Mistral AI
7 models
Model IDContextBest ForTag
mistral-large-3128kGeneral + toolsTOOLS
mistral-small-3.1128kVision + smallVISION
mistral-nemo-instruct-2407128k12B general12B
mistral-7b-instruct-v0.332kLightweight7B
mistral-codestral-2501256kCode specialistCode
mistral-codestral-mamba-v0.1256kCode Mamba archCode
mistral-mixtral-8x7b-instruct32kMoE modelMoE
Model IDProviderContextBest For
kimi-k2.5MoonShot128kLong context, vision
perplexity-sonarPerplexity128kBuilt-in web search
perplexity-sonar-proPerplexity128kPro search quality
phi-4Microsoft16kSmall reasoning
phi-4-miniMicrosoft16kUltra lightweight
nova-liteAmazon300kFast + cheap
nova-microAmazon128kUltra fast
nova-proAmazon300kMost capable Nova
cloudflare-llama-3.1-8bCloudflare128kEdge CDN inference
cloudflare-qwen-1.5-7bCloudflare32kEdge Qwen
cloudflare-mistral-7bCloudflare32kEdge Mistral
Vision — Image Input Capable
16 models
Model IDProviderBest For
gemini-3-flash-previewGoogle★ Best vision overall
gemini-2.5-flashGoogleFast vision
gemini-2.5-proGoogleDetailed vision
gemini-1.5-flashGoogleQuick image analysis
gemini-1.5-proGoogleLong image context
gpt-4oOpenAIVision + reasoning
gpt-4o-miniOpenAICheap vision
gpt-4.1OpenAILatest vision
grok-3-vision-betaxAICreative + vision
kimi-k2.5MoonShotVision + long context
llama-3.2-11b-vision-instructMetaOpen source vision
llama-3.2-90b-vision-instructMetaLarge open vision
llama-4-maverickMetaLatest multimodal
mistral-small-3.1MistralVision + tools
mistral-large-3MistralVision analysis
claude-3-5-sonnet-20241022AnthropicVision + writing
Pass as model in action: "generate_image". ⚠ Images expire in 2 hours!
zimage ★
Versatile, balanced quality — BEST overall
512px: $0.005
1024px: $0.020
flux
Photorealistic, general purpose, cheapest
512px: $0.003
1024px: $0.012
klein
Artistic, stylized, creative renders
512px: $0.006
1024px: $0.024
klein-large
High-detail art, professional quality
512px: $0.012
1024px: $0.048
gptimage
AI-assisted, conceptual generation
512px: $0.040
1024px: $0.160
📡 All API Actions
Send action in JSON POST body to https://osint.lostingness.site/nexusai.php
nexusv1 nexusV1 — Smart Auto-Routing Best model selected automatically · web search · memory · ★ RECOMMENDED
ParameterTypeRequiredDescription
action string nexusv1
messages array [{role,content}] conversation history
system string Custom persona/instruction
temperature float 0.0–2.0, default: 0.7
max_tokens int Max tokens, default: 1200
multi_model bool Multi-model synthesis (3 API calls)
image_url string Image URL or base64 for vision
Response key: content
chat Chat Completions Direct model access · OpenAI-compatible format
ParameterTypeRequiredDescription
action string chat
model string Any model ID from models list
messages array Messages array
system string System prompt
temperature float 0.0–2.0
max_tokens int Max tokens
Response key: content
generate_image Image Generation 5 models · up to 2048×2048 · zimage recommended
ParameterTypeRequiredDescription
action string generate_image (or "image", "img")
prompt string Detailed image description
model string zimage★, flux, klein, klein-large, gptimage
width int Width px, max 2048 (default: 512)
height int Height px, max 2048 (default: 512)
Response key: image_url
vision Vision / OCR Image analysis · URL or base64 · OCR · multimodal
ParameterTypeRequiredDescription
action string vision (or "analyze_image", "ocr")
image_url string HTTPS URL or data:image/jpeg;base64,...
question string Question about image (default: Describe this)
model string Vision model (default: best available)
max_tokens int Max tokens (default: 1024)
Response key: reply
text Text Generate (Legacy) Simple prompt → response, no message array needed
ParameterTypeRequiredDescription
action string text (or "text_generate", "generate")
model string Model ID
prompt string Plain text prompt
system string System instruction
temperature float 0.0–2.0
max_tokens int Max tokens (default: 500)
Response key: completion
responses Responses API — Stateful Chained conversations · Nexus manages history server-side
ParameterTypeRequiredDescription
action string responses
model string Model ID
input string Your message
instructions string System-level instructions
previous_response_id string Chain to previous turn
Response key: text + response_id
models List All Models Get all available models including nexusV1 variants
ParameterTypeRequiredDescription
action string models (or "list_models")
Response key: data[] array
memory_view Memory Operations View/clear/save AI memory (persistent Q&A store)
ParameterTypeRequiredDescription
action string memory_view | memory_clear | memory_save
question string * For memory_save: question text
answer string * For memory_save: answer text
Response key: memory[], count
search Web Search Test Test Tavily + Serper search directly, see what context gets injected
ParameterTypeRequiredDescription
action string search (or "web_search")
query string Search query
Response key: tavily{}, serper{}
detect_search Detect Search Need Check if a query triggers auto web search
ParameterTypeRequiredDescription
action string detect_search
query string Query to analyze
Response key: result{search,conf,kw}, type
ping Ping / Health Check Verify API is online, check PHP version and memory count
ParameterTypeRequiredDescription
action string ping (or "health")
Response key: service, version, memory, php
⚡ Rate Limits
TypeLimitPerHTTP Response
Per minute20 requestsPer IP429 + Retry-After: 60
Per hour200 requestsPer IP429 + Retry-After: 3600
Per day800 requestsPer IP429 + Retry-After: 86400
Memory store200 Q&A entriesGlobalAuto-pruned
Image expiry2 hoursPer imageSave immediately!
🎮 API Playground
Interactive AI — chat, generate images, analyze with vision. All live.
⚙ Chat Config
nexusV1
Hello! I'm nexusV1, built by @lostingness. I auto-detect your question type and use the best AI model. I can search the web, remember our conversation, write code, analyze images, and much more. What would you like to know? 🚀
nexusV1 is thinking...
🔍 Web search will trigger
🖼 Image Config
Flux
$0.003
ZImage
★ Best
Klein
$0.006
Klein+
$0.012
GPTImg
$0.040
512²
768²
1024²
1024×576
576×1024
Generating...
Generated image
Model:
Size:
Expires: 2 hrs
↗ Open Full
🖼
Generated image will appear here
⚠ Images expire in 2 hours — download immediately!
👁 Vision Config
📁
Click or drag & drop image
Analyzing...
📝 Vision Response
Vision analysis will appear here...
🧪 API Tester
Send raw JSON requests to the API and see exact responses. Perfect for debugging integrations.
📤 Request
Quick Tests
📥 Response
Click "Send Request" to see the API response here...
👨‍💻
Mohit Gupta
@lostingness
Developer · AI Builder · Class 11 Commerce
Self-taught developer from Jamnagar, Gujarat. Building AI tools, automation bots, and web apps. Vibe Coder.
94+
AI Models
500+
Users Served
v6.0.0
API Version
"I'm a self-taught developer from Jamnagar, Gujarat building AI tools that actually work. lostingness AI is my way of making powerful AI accessible to everyone — free, fast, and built with passion."
🛠️ Built With
🐘
BACKEND
PHP 7.2+
Single-file API, cPanel ready
🔍
SEARCH 1
Tavily AI
Advanced web search + direct answers
🌐
SEARCH 2
Serper (Google)
Google results, knowledge graph, news
🤖
AI MODELS
nexusV1 Gateway
94+ models via unified API
🧠
MEMORY
JSON Memory Store
File-based persistent Q&A memory
🔒
SECURITY
Rate Limiting
Per-IP: 20/min, 200/hr, 800/day