§ developer API · public · CORS open
Build on atomeons.com.
Six public endpoints. Semantic Q&A, fuzzy search, embeddings, markdown export, sales counters. CORS open. CC-BY 4.0. No auth, no quotas, no API key, no signup.
POST
/api/ask
Semantic Q&A over atomeons.com's 256 routes. Returns a 2-5 sentence answer with route-level citations.
example
curl -X POST https://atomeons.com/api/ask \
-H 'content-type: application/json' \
-d '{"query":"what is prompt injection","k":5}'response · sample
{
"ok": true,
"mode": "fuzzy",
"query": "what is prompt injection",
"answer": "Prompt injection is ... (atomeons.com/q/what-is-prompt-injection)",
"sources": [
{ "route": "/q/what-is-prompt-injection", "title": "What Is Prompt Injection?",
"section": "main", "similarity": 0.53 },
...
],
"index_built": "2026-06-05",
"index_count": 247
}notes
Layer A is fuzzy keyword retrieval; Layer B (vector retrieval, gemini-embedding-001 768-dim) auto-promotes once /public/vector-index.json lands. Synthesis uses gemini-2.5-flash. Query max 600 chars · k range 1-10.
GET
/api/ask
Health check + index state.
example
curl https://atomeons.com/api/ask
response · sample
{
"ok": true,
"mode": "fuzzy",
"fuzzy_index": { "count": 247, "built": "2026-06-05" },
"vector_index": null,
"generate_model": "gemini-2.5-flash"
}GET
/api/search?q=…&k=10
Fuzzy keyword search · returns matching routes WITHOUT synthesis.
example
curl 'https://atomeons.com/api/search?q=prompt+injection&k=5'
response · sample
{
"ok": true,
"query": "prompt injection",
"count": 5,
"results": [
{
"route": "/q/what-is-prompt-injection",
"title": "What Is Prompt Injection?",
"description": "...",
"score": 23,
"headings": ["The short answer", "The longer answer"],
"snippet": "..."
},
...
]
}notes
CORS open · k range 1-50 · cached 60s edge.
POST
/api/embed
Gemini-embedding-001 vectors at 768 dimensions (Matryoshka). Single text or batch up to 32.
example
curl -X POST https://atomeons.com/api/embed \
-H 'content-type: application/json' \
-d '{"text":"hello world"}'response · sample
{
"ok": true,
"model": "gemini-embedding-001",
"dim": 768,
"embedding": [0.0123, -0.0456, ...]
}notes
Rate-limited upstream by Gemini free tier (5 RPM). Each text max 8000 chars. Batch endpoint accepts {texts: [...]} up to 32 items. taskType options: RETRIEVAL_DOCUMENT (default) · RETRIEVAL_QUERY · SEMANTIC_SIMILARITY · CLASSIFICATION.
GET
/api/md?route=/orangebox
Per-page markdown export of any indexed route. Cleaner for LLM ingestion than rendered HTML.
example
curl 'https://atomeons.com/api/md?route=/orangebox'
response · sample
# ORANGEBOX Version 1\n\n**Route:** `atomeons.com/orangebox`\n**License:** CC-BY 4.0\n\n## Description\n...\n\n## Body\n...
notes
404 returns close-route suggestions in markdown.
GET
/api/sales-count
Live sales counter + revenue + current dynamic price.
example
curl https://atomeons.com/api/sales-count
response · sample
{
"ok": true,
"ts": "2026-06-05T23:00:00Z",
"total_sales": 0,
"total_revenue_usd": 0,
"net_buyers": 0,
"current_price_usd": 1,
...
}§ machine-readable resources
- https://atomeons.com/openapi.json
OpenAPI 3.1 spec · machine-readable
- https://atomeons.com/.well-known/agent.json
Agent discovery manifest · endpoints + resources
- https://atomeons.com/.well-known/mcp.json
Model Context Protocol manifest · ask_atomeons tool
- https://atomeons.com/.well-known/ai-plugin.json
Legacy ChatGPT plugin manifest
- https://atomeons.com/llms.txt
Lab-content overview for LLM ingestion
- https://atomeons.com/llms.md
Same content, .md extension
- https://atomeons.com/llms-full.txt
Full corpus dump · regenerated per request · 15-min edge cache
- https://atomeons.com/search-index.json
Fuzzy search index · 247 routes · same data /api/search returns
- https://atomeons.com/llm-routes.json
Top 31 routes with JSON-LD + OG metadata
- https://atomeons.com/sitemap.xml
Global sitemap · every route
- https://atomeons.com/sitemap-ai.xml
AI-priority sitemap · 55 routes ranked
- https://atomeons.com/sitemap-news.xml
Founder's View nightly broadcast news sitemap
- https://atomeons.com/robots.txt
Crawler hints · 30+ AI bots explicitly allowed
§ if you build something with these
Email atom@atomeons.com. I want to see it. If it's interesting, I'll feature it in /now and probably write about it in Founder's View. CC-BY 4.0 means attribution preserved — that's the only ask.