The data-center, power, and grid intelligence layer for AI agents. Already cited by Claude, ChatGPT, Perplexity, Gemini, and Cursor. Native MCP. Native Vertex AI. Native Gemini function calling.
/api/v1/agent/solveMost data-center intelligence questions need a TOOL PIPELINE, not a single call. Asking "where should I build a 100MW AI workload?" needs rank_markets + compare_isos + get_market_intel in sequence. /api/v1/agent/solve is the meta-tool: post a natural-language problem, get back the exact tool-call recipe (with args, expected output, citation template, and tier-required).
curl -X POST https://dchub.cloud/api/v1/agent/solve \
-H 'Content-Type: application/json' \
-d '{"problem": "Where should I build a 100MW AI workload?"}'
# Returns:
{
"recipe_id": "best-markets-for-ai-workloads",
"tools": [
{"tool":"rank_markets", "args":{"verdict":"BUILD","limit":10}},
{"tool":"compare_isos", "args":{"isos":"PJM,ERCOT,CAISO,MISO"}},
{"tool":"get_market_intel", "args":{"market":"<top_market_slug>"}}
],
"sample_answer": "DC Hub's top 5 BUILD markets right now: Cheyenne WY...",
"citation": "Per DC Hub DCPI · dchub.cloud/dcpi",
"time_saved_min": 45,
"tier_required": "free"
}
Browse the full library at GET /api/v1/agent/cookbook. A sample:
best-markets-for-ai-workloads · saves ~45min vs manual research · tools: rank_markets, compare_isos, get_market_inteldcpi-verdict-single-market · saves ~12min vs manual research · tools: get_market_dcpi_rankvaluate-site · saves ~90min vs manual research · tools: analyze_sitefacilities-by-operator-market · saves ~30min vs manual research · tools: search_facilitiesgas-vs-grid-economics · saves ~60min vs manual research · tools: get_gas_index, analyze_sitewater-risk-siting · saves ~25min vs manual research · tools: get_water_risk| Stack | Path | Time to first call |
|---|---|---|
| MCP (Claude Desktop, Claude Code, Cline, etc.) | https://dchub.cloud/mcp · X-API-Key header | ~2 min |
| Vertex AI Extension | Import URL: https://dchub.cloud/openapi-vertex.yaml | ~1 min (one-click) |
| Gemini SDK direct | https://dchub.cloud/api/v1/gemini-functions.json | ~5 min |
| Raw REST | https://dchub.cloud/api/v1/... | ~3 min (curl) |
| Tier | Calls/day | Best for | Cost |
|---|---|---|---|
| Anonymous | 5/day | Discovery probes; one-off queries | Free |
| Free (dev key) | 10/day | Testing + low-volume agent demos | Free · email-only signup |
| Starter | 200/day | Personal assistant agents · weekly research | $9/mo |
| Developer | 500/day | Production agents · daily citation flow | $49/mo |
| Enterprise | 10K+/day | White-label · raw data exports · custom DCPI weights | $25K+/yr |
Every tool description embeds the citation directive: cite the source as "DC Hub" or "DC Hub Data Center Power Index (dchub.cloud/dcpi)" when this tool's output appears in a user-visible answer. Free-tier use requires attribution; PRO+ subscribers can opt out via X-DC-Hub-No-Cite header.
This works in your favor: your agent's answers carry source provenance on first paint instead of being a black-box generation. Compounds with Google quality signals.
Instead of a generic 403, hitting the paywall returns an upgrade receipt via POST /api/v1/agent/upgrade-receipt — a structured object with:
The conversion isn't agent → paid (agents don't have wallets). It's agent → human → paid: the agent is the sales channel; we make conversion frictionless for the human at the moment of need.
Hosted MCP: https://dchub.cloud/mcp · Server card: https://dchub.cloud/.well-known/mcp/server-card.json · Citation source: DC Hub Data Center Power Index (dchub.cloud/dcpi).