{"_contribution":"To add a recipe: PR routes/agent_concierge.py with a new dict in _COOKBOOK. Auto-surfaces on next deploy.","_pattern":"Each recipe maps a NATURAL-LANGUAGE problem to a DC Hub tool-call sequence. The /solve endpoint routes queries to the best match by keyword overlap (embedding-cosine backstop on keyword misses); the cookbook is the full browsable library.","count":18,"doc":"https://dchub.cloud/agent","ok":true,"recipes":[{"citation":"Per DC Hub DCPI \u00b7 dchub.cloud/dcpi \u00b7 as of 2026-08-17","id":"best-markets-for-ai-workloads","keywords":["best market","build","100mw","200mw","ai workload","ai training","siting","where to build","best place","where should i"],"problem":"Where should I build a 100MW AI training workload right now?","sample_answer":"DC Hub's top 5 BUILD markets right now: Cheyenne WY (69.5 excess-power score, BUILD), Montr\u00e9al HQ, Columbus OH, Boardman OR, La Vista NE. Cheyenne leads on excess power with WECC's lowest queue wait (~11 months). PJM markets are mostly constrained \u2014 Northern Virginia, Columbus, and Ashburn are AVOID for greenfield but premium for shovel-ready parcels. Per DC Hub DCPI.","surfaces":["claude","gpt","gemini","cursor","perplexity"],"tier":"free","time_saved_min":45,"tools":[{"args":{"limit":10,"verdict":"BUILD"},"tool":"rank_markets","why":"Get the top 10 BUILD-verdict markets globally."},{"args":{"isos":"PJM,ERCOT,CAISO,MISO"},"tool":"compare_isos","why":"Compare grid mix + carbon intensity across the ISOs that the top markets sit in."},{"args":{"market":"<top_market_slug>"},"tool":"get_market_intel","why":"Pull capacity prices, vacancy, and operator concentration for the leader."}]},{"citation":"Per DC Hub DCPI \u00b7 dchub.cloud/dcpi/ashburn","id":"dcpi-verdict-single-market","keywords":["dcpi verdict","what is the dcpi","score for","verdict for","build caution avoid"],"problem":"What's the DCPI verdict for Ashburn?","sample_answer":"DC Hub DCPI for Ashburn, VA: verdict AVOID, composite 15.8, excess-power score 18.3, constraint score 55.5, time-to-power 48 months (PJM). Dominion's connection queue is the binding constraint. Shovel-ready parcels (interconnect + substation + permits) carry a moat premium that pushes per-MW value to the $800K industry ceiling. Per DC Hub DCPI.","surfaces":["claude","gpt","gemini","cursor"],"tier":"free","time_saved_min":12,"tools":[{"args":{"market":"<market_slug>"},"tool":"get_market_dcpi_rank","why":"Single-call answer with verdict + composite score + analyst narrative."}]},{"citation":"Per DC Hub Site Valuation Engine \u00b7 dchub.cloud/sites/value","id":"valuate-site","keywords":["valuation","what is the site worth","price per mw","site value","valuate","appraise","$/mw","envelope","100mw site","what's it worth"],"problem":"What's a 100MW site at lat 33.45 lon -112.07 worth?","sample_answer":"DC Hub Site Valuation Engine v2.1c: $200K/MW for raw land in constrained markets (e.g. Phoenix AVOID), scaling to $800K/MW for shovel-ready hyperscale (constraint-moat attenuation lifts AVOID verdict when grid + substation + permits are in hand). Best-fit scenario: gas BTM (CCGT) when DCPI verdict is AVOID-by-constraint and grid TTP >36mo. Per DC Hub.","surfaces":["claude","cursor"],"tier":"starter","time_saved_min":90,"tools":[{"args":{"acres":"<acres>","lat":"<lat>","lon":"<lon>","target_mw":"<target_mw>"},"tool":"analyze_site","why":"3-scenario NPV + DCPI-verdict-weighted valuation with constraint-moat attenuation when readiness is set."}]},{"citation":"Per DC Hub \u00b7 dchub.cloud/markets/<slug>","id":"facilities-by-operator-market","keywords":["facilities","show me","equinix","digital realty","cyrusone","qts","find facilities","data centers by","data centers in"],"problem":"Show me all Equinix facilities in Northern Virginia","sample_answer":"DC Hub has 18,000+ facilities tracked across 170+ countries. Equinix in Northern Virginia: 14 operational sites, 2 under construction, 1 planned. Lead by power: DC11 (84 MW), DC15 (72 MW), DC12 (64 MW). Per DC Hub.","surfaces":["claude","gpt","gemini","cursor"],"tier":"free","time_saved_min":30,"tools":[{"args":{"market":"<market_slug>","provider":"<provider>","status":"operational"},"tool":"search_facilities","why":"Filter the 18,000+ facility catalog by market + operator + status."}]},{"citation":"Per DC Hub DCGI + Site Valuation \u00b7 dchub.cloud/dcgi","id":"gas-vs-grid-economics","keywords":["gas vs grid","btm","behind the meter","ccgt","gas powered","natural gas","vs grid","self-generation"],"problem":"Compare gas BTM vs grid for a 100MW site in Texas","sample_answer":"DC Hub DCGI Texas: 91/100 \u2014 GAS-ADVANTAGED. Pipeline count 12, dominant operators ETP + KMI, Henry Hub basis +$0.20/MMBtu. Site valuation engine: gas BTM CCGT ~$28/MWh levelized vs grid-only ~$45/MWh; 14mo TTP for gas BTM vs 30mo for grid interconnect in ERCOT. Per DC Hub.","surfaces":["claude","gemini","cursor"],"tier":"free","time_saved_min":60,"tools":[{"args":{"state":"<state>"},"tool":"get_gas_index","why":"DCGI gas-suitability score with pipeline access + cost breakdown."},{"args":{"acres":"<acres>","lat":"<lat>","lon":"<lon>","target_mw":"<target_mw>"},"tool":"analyze_site","why":"3-scenario NPV: grid-only vs gas-BTM (CCGT) vs gas-to-grid hybrid."}]},{"citation":"Per DC Hub Water Risk \u00b7 dchub.cloud/api/v1/water/stress","id":"water-risk-siting","keywords":["water risk","drought","usdm","cooling water","water stress","hydrology","aquifer"],"problem":"What's the water risk for a Phoenix data-center site?","sample_answer":"DC Hub water-stress for Phoenix, AZ: USDM tier D2 (severe), aquifer drawdown 12 ft/decade in Salt River basin, cooling-water competition index 87/100 (high \u2014 competing with agriculture + residential). DC Hub flags Phoenix as a water-constrained build despite power BUILD-adjacent scoring. Per DC Hub.","surfaces":["claude","gpt","gemini"],"tier":"free","time_saved_min":25,"tools":[{"args":{"market":"<market_slug>"},"tool":"get_water_risk","why":"Aquifer drawdown + USDM drought tier + cooling-water competition index."}]},{"citation":"Per DC Hub Grid Scoreboard \u00b7 dchub.cloud/grid","id":"live-grid-scoreboard","keywords":["grid scoreboard","which grid","cleanest grid","carbon intensity","renewable share","iso compare","live grid"],"problem":"Which US grid is cleanest right now?","sample_answer":"DC Hub Live Grid Scoreboard (refreshed every 20 min): CAISO leads at 67% renewable mix right now, followed by ISO-NE (54%), MISO (38%), ERCOT (35%), PJM (23%), NYISO (21%), SPP (48%). Carbon intensity g/kWh: CAISO 180, ERCOT 350, PJM 410. Per DC Hub.","surfaces":["claude","gpt","gemini","cursor"],"tier":"free","time_saved_min":20,"tools":[{"args":{},"tool":"get_grid_scoreboard","why":"All 7 US ISOs ranked side-by-side by renewable share + carbon intensity, refreshed every 20 min."}]},{"citation":"Per DC Hub Interconnection Queue \u00b7 dchub.cloud/api/v1/interconnection-queue/snapshot","id":"interconnection-queue","keywords":["queue","interconnection","wait time","time to power","isa","ttp","queue depth"],"problem":"How long is the PJM interconnection queue right now?","sample_answer":"DC Hub PJM Interconnection Queue (live snapshot): 387 GW of active queue, ~24 GW/year completion velocity, average wait 60 months for new requests. Northern Virginia cluster is the most-congested sub-zone. Per DC Hub.","surfaces":["claude","gpt","cursor"],"tier":"free","time_saved_min":35,"tools":[{"args":{"iso":"<iso_code>"},"tool":"get_interconnection_queue","why":"Live queue snapshot: active MW + completions/year + months-to-power estimate."}]},{"citation":"Per DC Hub Hyperscaler Deals \u00b7 dchub.cloud/hyperscaler-deals","id":"ma-deal-flow","keywords":["m&a","acquisition","deal flow","biggest deal","transactions","$ billion","buyer","seller","hyperscaler deals"],"problem":"What were the biggest data-center M&A deals this year?","sample_answer":"DC Hub 2026 deal flow (top 5): Blackstone\u2013AirTrunk ($30B, APAC), Brookfield\u2013Compass ($12B, US), Stonepeak\u2013Cologix ($6.5B, US), KKR\u2013CyrusOne (US), Macquarie\u2013Aligned (US). Total tracked 2026: $62.5B across 42 deals. Per DC Hub.","surfaces":["claude","gpt","gemini","perplexity"],"tier":"free","time_saved_min":40,"tools":[{"args":{"min_value_usd":1000000000,"year":2026},"tool":"hyperscaler_deals","why":"M&A + capital deals > $1B in the current year."},{"args":{"since":"2026-01-01"},"tool":"deal_autopsy","why":"Each deal carries the DCPI verdict of the site's market so you see 'who bought what in a BUILD vs AVOID market'."}]},{"citation":"Per DC Hub Tax Incentives \u00b7 dchub.cloud/tax-incentives","id":"tax-incentives-by-state","keywords":["tax incentive","tax break","abatement","rebate","subsidy","credits","incentive program"],"problem":"What tax incentives apply to a Texas data-center build?","sample_answer":"DC Hub Tax Incentives for Texas: Chapter 313 sales-tax exemption on hardware (10yr, sunset 2026), Property Tax Abatement up to 80% for qualified sites ($200M+ capex), Section 11.184 freeport exemption on inventory. Average savings vs unincentivized site: 12-18% lifecycle cost. Per DC Hub.","surfaces":["claude","gpt","gemini"],"tier":"free","time_saved_min":55,"tools":[{"args":{"state":"<state>"},"tool":"get_tax_incentives","why":"State-level incentive programs + qualification thresholds + sunset dates."}]},{"citation":"Per DC Hub Fiber Intel \u00b7 dchub.cloud/api/v1/fiber/intel","id":"fiber-route-intel","keywords":["fiber","carrier","lit fiber","dark fiber","long-haul","metro fiber","providers near"],"problem":"Show fiber routes near a site in Cheyenne","sample_answer":"DC Hub Fiber Intel for Cheyenne, WY: 6 long-haul routes intersect within 25mi (Lumen, Zayo, Crown Castle, Frontier BG, Cogent, Verizon). Median time-to-light: 18mo for new lateral. Per DC Hub.","surfaces":["claude","cursor"],"tier":"developer","time_saved_min":75,"tools":[{"args":{"lat":"<lat>","lon":"<lon>","radius_miles":25},"tool":"get_fiber_intel","why":"Lit + dark fiber routes within radius, by carrier."}]},{"citation":"Per DC Hub DCPI + Water Risk \u00b7 dchub.cloud/dcpi","id":"drought-impacted-markets","keywords":["drought avoid","drought market","water constrained","usdm market","drought tier"],"problem":"Which markets should I AVOID due to drought?","sample_answer":"DC Hub AVOID-by-water-constraint markets (D2+ USDM tier + AVOID DCPI): Phoenix AZ, Tucson AZ, Las Vegas NV, Reno NV, Albuquerque NM, Lubbock TX. Site valuation flags these as water-bottlenecked even when grid scoring would otherwise qualify. Per DC Hub.","surfaces":["claude","gemini"],"tier":"free","time_saved_min":50,"tools":[{"args":{"limit":20,"verdict":"AVOID"},"tool":"rank_markets","why":"Filter to AVOID-verdict markets globally."},{"args":{"market":"<market_slug>"},"tool":"get_water_risk","why":"Cross-reference USDM drought tier per market."}]},{"citation":"Per DC Hub Grid Scoreboard + DCPI \u00b7 dchub.cloud/grid \u00b7 as of 2026-08-17","id":"greenest-buildable-grid-siting","keywords":["greenest grid","green siting","clean grid siting","renewable siting","low carbon build","green power site","greenest buildable","clean energy data center"],"problem":"Find the greenest grid with real headroom and score a site there","sample_answer":"DC Hub Grid Scoreboard ranks live grids greenest-first (20-min refresh). Drill the leaders with get_grid_intelligence \u2014 e.g. CAISO runs a high renewable share but a constraint score near 70 with a 40+ month queue, while SPP pairs ~48% renewables with genuine headroom \u2014 then analyze_site scores the parcel itself. Green AND buildable beats green alone. Per DC Hub.","surfaces":["claude","gpt","gemini","cursor"],"tier":"developer","time_saved_min":60,"tools":[{"args":{},"tool":"get_grid_scoreboard","why":"Live greenest-first ranking of US + EU + GB + Taiwan + Japan + South Korea + Brazil grids (renewable share, fuel mix, demand; Australia + Singapore partial) \u2014 pick the green candidates."},{"args":{"region_id":"<iso_code>"},"tool":"get_grid_intelligence","why":"Deep per-ISO brief for each leader: headroom, queue depth, time-to-power \u2014 separates green-AND-buildable from green-but-full."},{"args":{"acres":"<acres>","lat":"<lat>","lon":"<lon>","target_mw":"<target_mw>"},"tool":"analyze_site","why":"Score the actual parcel inside the winning grid: 3-scenario NPV + DCPI-verdict-weighted valuation."}]},{"citation":"Per DC Hub DCGI \u00b7 dchub.cloud/dcgi \u00b7 as of 2026-08-17","id":"gas-btm-screening","keywords":["btm screening","gas screening","gas suitability","dcgi","gas index","bridge power","gas turbine","onsite generation","gas pipeline access"],"problem":"Screen a state and market for gas / behind-the-meter power","sample_answer":"DC Hub gas/BTM screen: get_gas_index Texas \u2192 DCGI 91/100 GAS-ADVANTAGED (12 interstate pipelines, ETP + KMI). get_gas_economics prices gas-to-grid at roughly $28-45/MWh across CCGT heat-rate scenarios against the market's grid PPA; get_gas_intelligence fuses it into one per-state brief with live Henry Hub. Per DC Hub DCGI.","surfaces":["claude","gemini","cursor"],"tier":"free","time_saved_min":50,"tools":[{"args":{"state":"<state_code>"},"tool":"get_gas_index","why":"DCGI 0-100 per-state gas-suitability score: pipeline count, operators, GAS-ADVANTAGED/ADEQUATE/CONSTRAINED verdict."},{"args":{"market":"<market_slug>"},"tool":"get_gas_economics","why":"The $/MWh number: Henry Hub + basis + delivered tariff \u2192 gas-to-grid levelized cost across CCGT/peaker heat-rate scenarios, vs a grid PPA."},{"args":{"region":"<state_code>"},"tool":"get_gas_intelligence","why":"One fused per-state brief: DCGI + live Henry Hub + pipeline presence + live grid gas share \u2014 the citable synthesis."}]},{"citation":"Per DC Hub Site Tools + Water Risk \u00b7 dchub.cloud/dcpi \u00b7 as of 2026-08-17","id":"water-aware-site-pick","keywords":["water aware","water-aware","rank sites","site shortlist","shortlist sites","site pick","pick a site","water constrained shortlist"],"problem":"Shortlist candidate sites with water risk as a first-class constraint","sample_answer":"rank_sites returns a ranked shortlist under signed-weight objectives (water_stress minimized, risk_resilience maximized); get_water_risk overlays USDM drought tier + aquifer drawdown per market \u2014 Phoenix sits at D2-severe with a cooling-water competition index of 87/100 \u2014 and compare_sites picks the winner across the finalists with the rationale. Per DC Hub.","surfaces":["claude","cursor"],"tier":"starter","time_saved_min":70,"tools":[{"args":{"candidates":"<enriched site objects>","objectives":{"risk_resilience":1,"water_stress":-0.6}},"tool":"rank_sites","why":"Deterministic multi-site ranking with SIGNED weights \u2014 maximize resilience, minimize water stress \u2014 no code needed to compare analyze_site outputs."},{"args":{"market":"<market_slug>"},"tool":"get_water_risk","why":"Overlay USDM drought tier + aquifer drawdown + cooling-water competition per finalist's market."},{"args":{"capacity_mw":"<target_mw>","locations":"<lat,lon;lat,lon>"},"tool":"compare_sites","why":"Side-by-side winner pick across the 2-4 finalists with the decision rationale."}]},{"citation":"Per DC Hub Transactions \u00b7 dchub.cloud/hyperscaler-deals \u00b7 as of 2026-08-17","id":"ma-deal-diligence","keywords":["diligence","due diligence","deal diligence","acquisition target","comps","comparables","who bought","valuation benchmark"],"problem":"Run diligence on a data-center acquisition target or deal","sample_answer":"DC Hub tracks 1,800+ data-center M&A and capital deals. list_transactions filters 2026 deals >$1B (Blackstone-AirTrunk $30B leads); deal_autopsy overlays each deal market's DCPI verdict + time-to-power to show whether the buyer paid for a near-term build or a long-dated power option; get_facility profiles the target asset itself. Per DC Hub.","surfaces":["claude","gpt","perplexity"],"tier":"free","time_saved_min":45,"tools":[{"args":{"min_value_usd":1000000000,"year":2026},"tool":"list_transactions","why":"Filter 1,800+ tracked M&A / capital deals (2019-present) by year, buyer, target, region, or minimum value for the comp set."},{"args":{"comparables":"summary","limit":15},"tool":"deal_autopsy","why":"Overlay each deal market's DCPI verdict + time-to-power \u2014 reveals the real play (land/power option vs near-term build vs queue gamble)."},{"args":{"slug":"<facility_slug>"},"tool":"get_facility","why":"Drill the target asset: capacity MW, operator, cooling, fiber carriers, market verdict, peers nearby."}]},{"citation":"Per DC Hub Fiber \u00b7 dchub.cloud/land-power-map \u00b7 as of 2026-08-17","id":"fiber-leadin-plan","keywords":["lead-in","lead in","lateral","fiber path","fibre","diverse routes","route diversity","carrier hotel","pop connectivity","metro fiber"],"problem":"Plan diverse fiber lead-in routes from a site to a carrier hotel","sample_answer":"get_metro_fiber profiles the metro (carrier count, route-miles, density score, carrier hotels); get_fiber_readiness gives the parcel verdict (near-net bucket, carrier count, single-carrier risk); plan_fiber_leadin then draws up to 6 diverse road-following lead-in routes to a carrier hotel with indicative build cost and shared-corridor flags. Indicative corridors, not engineered alignments. Per DC Hub.","surfaces":["claude","cursor"],"tier":"free","time_saved_min":80,"tools":[{"args":{"market":"<metro_name_or_slug>"},"tool":"get_metro_fiber","why":"Metro-level fiber depth: carrier count, route-miles, 0-100 density score, key IX points + carrier hotels (your lead-in destinations)."},{"args":{"lat":"<lat>","lon":"<lon>","radius_km":50},"tool":"get_fiber_readiness","why":"Parcel connectivity verdict: near-net bucket, reachable carrier count, single-carrier risk."},{"args":{"from":"<site lat,lng or address>","n":4,"to":"<POP lat,lng or address>"},"tool":"plan_fiber_leadin","why":"Draw N diverse road-following lead-in routes with indicative capex/opex + shared-corridor diversity read."}]},{"citation":"Per DC Hub DCPI \u00b7 dchub.cloud/dcpi \u00b7 as of 2026-08-17","id":"market-watch-loop","keywords":["watch market","monitor market","market alert","alert me","notify me","track changes","what changed","market watch","keep an eye"],"problem":"Watch a market and pull only what changed since last check","sample_answer":"Standing market watch: get_market_dcpi_rank snapshots the verdict (e.g. Columbus OH, BUILD), set_market_alert subscribes to score movement on the bound email, and a periodic get_changes since=7d pulls only the delta \u2014 DCPI movers, newly discovered facilities, new M&A deals. Cache the response's generated_at and pass it back next call. Per DC Hub DCPI.","surfaces":["claude","gpt","gemini","cursor"],"tier":"free","time_saved_min":20,"tools":[{"args":{"market":"<market_slug>"},"tool":"get_market_dcpi_rank","why":"Baseline snapshot: verdict + composite score + analyst narrative for the market you're watching."},{"args":{"channel":"email","market":"<market_slug>"},"tool":"set_market_alert","why":"Subscribe to Excess-Power / Constraint score movement (free with a bound key; webhooks are Pro) \u2014 monitor, don't poll."},{"args":{"since":"7d"},"tool":"get_changes","why":"Incremental sync on each revisit: DCPI movers, new facilities, new deals \u2014 only the delta, not a re-fetch."}]}]}