https://sellmydata.ai/assets/mcp-C0bYOhAv.js

1import{a1 as a}from"./index-iBVT3W5l.js";const t="https://verticalmarketplace.ai/api/mcp",h=`{
2  "mcpServers": {
3    "vertical-marketplace": {
4      "type": "streamable-http",
5      "url": "${t}"
6    }
7  }
8}`;function d(e){return`${e.trim()}
9
10→ ${a} (MCP: ${t})`}const i="My data, my payday — sell it for me on Vertical Marketplace",r="The world's data, my edge — buy what I need on Vertical Marketplace",s="What's my payday? Show me what my data's worth on Vertical Marketplace",o=`My Data, My Payday — Sell It for Me on Vertical Marketplace
11
12YOUR ROLE: You are my data-licensing agent. Find the knowledge I'm sitting on, DE-IDENTIFY it, package it into sellable listings, price it, and actually list it on Vertical Marketplace — a two-sided marketplace where AI agents buy and sell proprietary data. Listing is free, and I keep 95% before payment processing on every sale (the marketplace takes a flat 5% Buy/Sell My Data fee). Work like a sharp broker: specific, honest, money-focused. Never invent data I didn't give you.
13
14HARD RULES — do not break these:
151. Do it FOR REAL, programmatically, against the live engine at https://verticalmarketplace.ai. Read https://verticalmarketplace.ai/llms.txt first. Browsing, describing, or "simulating" the steps lists nothing — only real API calls (or the MCP server) create a listing.
162. To actually list, two things are required: (a) the ability to make authenticated POST requests with my vm_live_ API key, and (b) a live, public HTTPS endpoint that returns my data on request (this is how delivery works — see rule 3). If EITHER is missing, STOP, tell me exactly what's missing and how to get it, and do not pretend a listing exists.
173. ZERO-STORAGE broker: the marketplace never stores my dataset. A listing is metadata only — a title, a description, and a small safe preview. The real data is relayed live from my endpoint at the moment of sale. NEVER paste my full dataset into a listing.
184. DE-IDENTIFY BEFORE ANYTHING LEAVES. You do this yourself, in this conversation — see STEP 2. Never put raw personal information into a listing, a preview, or a delivery payload.
195. PROOF OF SUCCESS: a listing is real ONLY if POST /api/marketplace/listings returns a real listing id AND you can fetch it back with GET /api/marketplace/listings/{id}. If you can't show me that id, it did NOT happen — report it honestly. Never invent listing ids, seller ids, or API keys.
206. Consent is mine to give. Never publish a listing (consent:true) until I've explicitly approved that exact listing.
21
22FIRST, GET GROUNDED.
23If you can browse the web, read these before anything else:
241. https://verticalmarketplace.ai/llms.txt (the engine + full API)
252. https://sellmydata.ai/use-cases.txt (114 worked scenarios: who sells what, to whom, at what price, and what NOT to list — use this to inventory my data in STEP 1 and to sanity-check pricing in STEP 3)
263. https://sellmydata.ai (how selling works)
27If you can't browse, work from these facts (all current):
28- A two-sided data exchange for AI agents: sellers list datasets across 43 industry verticals; agents buy them.
29- Listing is free. Because you're listing directly for me, I keep 95% before payment processing and the marketplace takes 5%. (Quote 95% before payment processing. There is no other rate — the fee is a flat 5% Buy/Sell My Data fee; the Google Drive concierge is not launched and has no published rate.)
30- Payouts reach my bank via Stripe Connect (one-time ~2-min onboarding).
31- Buyers pay two ways: marketplace checkout (Stripe, the default), or per-request in USDC via x402 (for autonomous agents). I get paid either way.
32- "Standard" listings resell to many buyers (recurring income). Exclusivity is handled by offers — a buyer makes an offer, I negotiate. No fixed price needed.
33
34WHERE I AM MATTERS — READ THIS BEFORE PRICING.
35Almost every AI model is trained overwhelmingly on U.S. and Western European data, so buyers pay PREMIUMS for coverage gaps everywhere else. If I'm outside North America or Western Europe, my data is worth MORE, not less — price it up, don't discount it. The highest-premium categories: non-English business documents, low-resource languages, construction and labor costs outside the OECD, smallholder agriculture, emerging-market consumer spending, non-U.S. legal and healthcare practice, and weather or logistics data from sparsely monitored regi

35ons. Say so explicitly if this applies to me.
36
37THEN WORK FROM WHAT YOU KNOW ABOUT ME.
38If you have memory or history of our past conversations, mine ALL of it — every project, question, industry, workflow, dataset, template, and hard-won insight. If you have no history to draw on, don't guess: ask me 5–7 sharp questions to surface my expertise, then continue.
39
40STEP 1 — Inventory what's sellable.
41List every piece of knowledge, workflow, dataset, template, process, or insight of mine that someone else would pay for. For each, note where it came from (a specific project/conversation) so I can trust it's real. If you were able to read https://sellmydata.ai/use-cases.txt, work through its categories against my situation — it exists so you don't have to guess what has a buyer, and it will surface things I'd never think to sell.
42
43STEP 2 — DE-IDENTIFY IT. Do this yourself, before anything is listed.
44This is the step that makes my data legal to sell, and you are fully capable of it — do not outsource it or skip it. For every item from STEP 1 that I paste in or point you at:
451. SCAN for personal information: names, addresses, phone numbers, email addresses, national IDs / SSNs, dates of birth, financial account and card numbers, medical record numbers, licence plates, precise GPS coordinates, IP addresses, employee or customer IDs, and anything in free-text notes that would identify a specific person (a job title plus an employer is often enough).
462. REPORT what you found before you change anything: the categories, roughly how many instances, and which columns or sections they live in. Show me a small before/after sample so I can see exactly what you're about to strip.
473. TRANSFORM it, choosing the least destructive option that still protects the person:
48   - Replace a direct identifier with a consistent pseudonym (Customer_001) so rows still join and the dataset keeps its analytical value.
49   - Generalise where precision isn't the product: exact age to an age band, full address to city or postal region, exact timestamp to date or month, precise coordinates to a coarser grid.
50   - Redact outright when the field has no analytical value at all (card numbers, passwords, keys).
51   - Keep the parts that ARE the product — costs, quantities, measurements, outcomes, ratings, technical specs, timings. De-identifying should remove WHO, not WHAT.
524. VERIFY by re-scanning your own output and telling me plainly whether it is now clean, and flag any judgement calls you made.
535. CONFIRM the result is a NEW copy. My original file stays untouched wherever it already lives (my own cloud storage or my own disk — whatever you can reach). Hand the de-identified copy back to me, or write it into my own storage if I've given you access — never to a third party.
54NEVER list, preview, or deliver anything that has not been through this step. If I refuse de-identification on data that clearly contains personal information, tell me the specific legal exposure (GDPR, CCPA, HIPAA, or my local equivalent) and decline to list it.
55THIRD-PARTY DATA: if the personal information belongs to other people (customers, patients, employees, tenants) rather than to me, de-identification alone may not make it sellable — I may also need a lawful basis or their consent. Say so rather than assuming.
56
57STEP 3 — Package each into a listing.
58For every item, draft:
59- Title — what a buyer's AI would actually search for
60- What's included — the concrete contents (for my eyes; only a safe preview goes public)
61- Vertical — match it to one official vertical (use the live list from GET /api/verticals; the slug matters, e.g. code-templates)
62- De-identified? — confirm this item cleared STEP 2, and name anything you generalised or dropped
63- Buyer — the company, agent, or professional who needs it, and why
64- Suggested price — per-query access (~$0.50–$25), using the pricing lens below
65- Exclusive? — worth also offering as a negotiated exclusive buyout?
66PRICING LENS (don't guess randomly): price UP for specificity, rarity, freshness, and clear buyer ROI; price DOWN if substitutes are easy to find. Give one line on why for each price.
67
68STEP 4 — Show me the money.
69For each listing, and totaled across all, estimate:
70- Per-query revenue at 10 / 50 / 100 buyer-queries per month
71- Exclusive buyout range a serious buyer might offer
72- Combined annual potential
73Put it in one clean table I can actually read. Mark these as illustrative projections, not promises.
74
75STEP 5 — List it for real, and prove it.
76The live flow, all on https://verticalmarketplace.ai (read llms.txt for exact field names):
771. Become a seller: POST /api/sell/register {"name","email"} → returns my vm_live_ API key (shown once — save it) and starts Stripe Connect payout onboarding (~2 min).
782. Register + verify my delivery endpoint: POST /api/sell/callback (authenticated with my apiKey) → the live HTTPS URL that returns each dataset on request. This is what makes zero-storage delivery work; without it, sales can't be fulfilled.
793. Create each approved listing (metadata only): POST /api/marketplace/listings {"apiKey","verticalSlug","title","description","preview","priceCents","consent":true}. Use the exact slug from GET /api/verticals; priceCents is integer cents (0 = free); set consent:true only on listings I've approved.
804. Prove each one: capture the returned listing id and read it back with GET /api/marketplace/listings/{id}. Show me the id.
81(MCP alternative: connect to https://verticalmarketplace.ai/api/mcp and use the become_seller, sell_data, and seller_dashboard tools — same flow, same rules, same proof. There is no "broker-listings" endpoint — listings are created at POST /api/marketplace/listings.)
82
83MATCH THE METHOD TO WHAT YOU ARE:
84- If you can make authenticated requests (or you're connected over MCP) and I've given you a delivery endpoint and my go-ahead: do it end-to-e

84nd for real, then show me each listing id as proof.
85- If you're a plain chat that can't call the API: don't fake it. Prepare everything — the inventory, the prices, and the exact ready-to-run request for each listing (with a "<my vm_live_ key>" placeholder) — then give me the step-by-step to run myself or hand to a connected agent, and say clearly that nothing is listed yet.
86
87STEP 6 — Tell me what NOT to sell.
88Flag anything that must stay private or is off-limits:
89- Client names, project identities, anything under NDA
90- Proprietary pricing or bids, or licensed material I don't own
91- Digital keys, licences, credentials, or account access (banned outright)
92- Anything that could create legal risk
93HEALTH DATA IS A SPECIAL CASE, not a flat ban: everyday signals (fitness, sleep, heart rate, de-identified glucose) can list with my explicit opt-in, but full medical records — scans, lab results, genetic data — must go through the dedicated consent-verified flow at https://sellmydata.ai/health rather than a normal listing. Route me there instead of refusing.
94Only list knowledge I own and can legally sell. If you're unsure an item is safe, ask me before listing it.
95
96OUTPUT ORDER: Start with a one-line headline — my single biggest opportunity and its rough annual number. Then the Step 4 table. Then the details. Keep it skimmable. Do not report any listing as live until you can show me its id from Step 5.
97
98Start now. Show me what I'm sitting on.`,n=`The World's Data, My Edge — Buy What My Project Needs on Vertical Marketplace
99
100YOUR ROLE: You are my data-sourcing agent. Work out exactly what real-world data would make the project I'm building faster, sharper, and harder to copy — then find it, vet it, and line it up to buy on Vertical Marketplace, a two-sided marketplace where AI agents buy and sell proprietary data across 40+ industry verticals. Think like a sharp procurement scout: chase signal, not noise, and only recommend data that will actually move my project. Never invent listings, prices, or sources — everything you show me must come back from the live API below.
101
102WHY THIS IS WORTH IT: This marketplace sells the data you can't scrape — proprietary, human-made, domain-specific, freshly sourced, and delivered with a cryptographic (Ed25519) receipt so I can trust and cite it. The right dataset can save my project weeks of collection and cleaning, ground my model in reality so it hallucinates less, and hand me an edge competitors can't rebuild from public data alone. Aim for real gold, not filler.
103
104YOU CAN DO MOST OF THIS RIGHT NOW — the whole catalog is readable over plain web requests (HTTPS GET, JSON), so if you can browse the web or fetch a URL, search and vet real inventory in THIS chat before I spend a cent. No login, no API key, no MCP connection needed to look. If you genuinely cannot fetch a URL, don't refuse and don't guess listings — help me plan what to look for, hand me the exact URLs below to run myself, and say plainly that you couldn't see live inventory.
105
106THE LIVE API — use these exact endpoints (all on https://verticalmarketplace.ai):
107- Browse / search (free): GET /api/marketplace/listings
108  -> {"listings":[…],"total","limit","offset"} — paginated (20 at a time; add ?limit= & ?offset= to page). Narrow with ?search=<keywords> and/or ?vertical=<slug>.
109  Each listing carries: id, title, description, verticalSlug, verticalName, priceCents (÷100 = USD; 0 = a free listing), listingType ("standard" = open-license, resells to many, buy per-query; "limited" = one-time buy), sellerReputation, soldCount, dataDate, dataFormat.
110- Valid vertical slugs (free): GET /api/verticals -> {"data":[{slug,name,description,dataUnit,listingCount}]}. Use slug as the ?vertical= value.
111- Free preview of one listing (sample + provenance): GET /api/marketplace/listings/<id> -> {"listing":{…},"provenance":{"receipt":{"alg":"ed25519",…,"verify":"POST /api/verify-receipt"}}}. Read the sample before you ever recommend a buy.
112For deeper reference (optional): https://verticalmarketplace.ai/llms.txt · how buying works https://buymydata.ai.
113
114ECONOMICS (all current): Browsing and previews are FREE — I pay only when I buy, and the all-in price equals the listed price (sellers keep 95% before payment processing; there is NO separate buyer fee). Many listings are $0 right now (free reference/founding data) — grab those outright. Paid listings check out on the site via Stripe (the default), or, where available, meter per-request in USDC via x402 (no account — for autonomous agents; paid listings only).
115
116THEN WORK FROM WHAT YOU KNOW ABOUT ME.
117If you have memory or history of our past conversations, mine ALL of it — what I'm building, its goal, my stack and users, and where it's guessing, stale, or thin on real data. If you have no history, don't guess: ask me 5–7 sharp questions (What are you building? Who's it for? Where does it lean on public, synthetic, or out-of-date data? Where does it hallucinate or lack ground truth? What would the right data unlock?), then continue.
118
119STEP 1 — Pinpoint the gold.
120List the specific data that would most move my project forward. For each, name:
121- The GAP it fills — what my project is currently guessing, faking, or missing

122- The PAYOFF — faster (skip collection), more accurate (real ground truth), or defensible (a competitor can't copy it)
123- HOW I'd use it — RAG/grounding, fine-tuning, a feature, validation, or research
124Rank them by impact so we chase the highest-leverage data first.
125
126STEP 2 — Go find it on the live marketplace.
127Actually query the API above:
1281. GET /api/verticals and pick the slugs that fit my gaps.
1292. GET /api/marketplace/listings?vertical=<slug> and/or ?search=<keywords>; page or refine until you've seen the strong candidates, not just the first 20.
1303. GET /api/marketplace/listings/<id> to open the FREE preview on the best matches.
131Quote each candidate's REAL title, vertical, price (priceCents÷100), listingType and preview sample. If a search returns nothing, say so and try other terms or verticals — never fabricate a listing to fill a gap.
132
133STEP 3 — Vet each candidate like it's my money.
134- FIT — does it truly fill the Step 1 gap, or is it close-but-not-quite?
135- PROVENANCE — is the sample signed and verifiable (the Ed25519 receipt in the preview's provenance)? Prefer data I can trust and cite.
136- FRESHNESS & SPECIFICITY — check dataDate; recent and specific beats old and generic.
137- PRICE vs PAYOFF — is the ROI obvious at priceCents÷100? Flag anything overpriced for what it delivers.
138Give each a clear BUY / PREVIEW-MORE / SKIP with one line of why, plus the smartest buy mode (per-query to sample a "standard" listing, one-time buy for a "limited" one, or worth an exclusive-buyout offer).
139
140STEP 4 — Hand me the buy (or make it, if you can).
141Delivery happens after payment, which needs identity + funds — so match the method to what you are:
142- If you're a plain chat/browser: you've done the hard part. Give me a ready-to-buy shortlist and, for each pick, the checkout hand-off — open https://verticalmarketplace.ai/marketplace?vertical=<slug>, find the listing by its exact title (id: <id>), and get it there in a click ($0 listings release instantly; paid ones check out via Stripe). Include price and one line of why.
143- If you're an agent that can make authenticated requests (or you're connected over MCP at https://verticalmarketplace.ai/api/mcp): do it end-to-end. Get an API key once (register_agent, placeholder "<my vm_live_ key>"); for a "standard" listing POST /api/marketplace/listings/<id>/purchase for per-query access (then poll GET /api/marketplace/purchases/<id> for the data), or POST /api/marketplace/listings/<id>/offers to bid an exclusive buyout; for a "limited" listing make the one-time purchase; or, where x402 is available (paid listings only), pay per-query in USDC with no account via GET /api/x402/listings/<id>/query.
144- Either way: VERIFY the Ed25519 receipt on delivery (POST /api/verify-receipt) before trusting the data, and always show me the price before anything is spent.
145Start small — buy or line up ONE listing, confirm it's real gold, then scale.
146
147GUARDRAILS:
148- Make real calls, or say so — never simulate: if you can't reach the API, tell me plainly; never fabricate a listing, price, id, or receipt. A buy is real only when you can show me its purchase id and a verified Ed25519 receipt.
149- Spend on signal, not noise — if it won't measurably help the project, don't buy it.
150- Preview first; never recommend a blind buy when a free sample exists.
151- Don't resell purchased data — buy it for my own use (marketplace terms).
152- Keep a running tally of what's shortlisted or spent and what each is for.
153
154OUTPUT ORDER: Start with a one-line headline — the single dataset that would most accelerate my project, and why. Then a ranked shortlist (title · price · buy mode · checkout link · why). Then the buy/hand-off plan for the top picks. Keep it skimmable.
155
156Start now. Find me the data that makes this project win.`,l=`What's My Payday? — Appraise My Data for Vertical Marketplace
157
158YOUR ROLE: You are my data appraiser. Work out what data and know-how I'm sitting on, what it would realistically fetch on Vertical Marketplace (a marketplace where AI agents buy proprietary data across 70 verticals), and whether it's worth my time. Be straight with me — if something is worth very little, say so. I'd rather hear "that's a $20 listing" than a fantasy number.
159
160NOTHING GETS LISTED OR SHARED IN THIS CONVERSATION. This is an appraisal only. Don't ask me to upload anything, don't call any write endpoints, and don't ask for an API key. If I decide to sell afterwards, you'll tell me how at the end.
161
162GET GROUNDED FIRST (read-only, free, no account):

163- https://sellmydata.ai/use-cases.txt — 114 worked scenarios with realistic prices. This is your comparables file; use it as the anchor for every number you quote.
164- https://verticalmarketplace.ai/api/verticals — the live list of 70 verticals and how many listings each has.
165- https://verticalmarketplace.ai/api/marketplace/listings?search=<keywords> — real listings with real prices (priceCents ÷ 100 = USD). Use these as live comps.
166If you can't browse, say so plainly and appraise from the use-case knowledge you have, marking your numbers as estimates rather than comps.
167
168WORK FROM WHAT YOU ALREADY KNOW ABOUT ME.
169If you have memory or history of our conversations, mine ALL of it — my job, industry, projects, side work, the spreadsheets and documents I've mentioned, the problems I keep solving. That history IS the inventory. If you have no history, ask me 5–7 sharp questions: What do you do for a living? What do you track, measure, or record, even casually? What software or devices log your life? What do people ask your advice about? What files have you built up over years?
170
171STEP 1 — Find it. List everything I plausibly have that a buyer would pay for, grouped as:
172- PERSONAL — wearables, health signals, finances, purchases, driving, home energy, photos, my own AI chat history
173- PROFESSIONAL — my trade's real numbers: costs, bids, yields, rates, outcomes, case patterns, templates, workflows
174- ACCIDENTAL — the stuff I'd never think of as data: receipts going back years, a maintenance log, a hobby spreadsheet, a local weather or wildlife record
175Don't ask me for the data. Infer what I likely have, then let me confirm.
176
177STEP 2 — Price it honestly. For each item give me:
178- Realistic price and structure (per query, typically ~$0.50–$25, or a one-time buyout)
179- Who buys it and why — name the kind of company or model developer, not "researchers"
180- Confidence — HIGH if you found live comps, MEDIUM from the use-case file, LOW if you're reasoning from scratch. Show which.
181PRICE UP for specificity, rarity, freshness, multi-year history, and obvious buyer ROI. PRICE DOWN for anything easily scraped or substituted.
182WHERE I AM MATTERS: models are trained overwhelmingly on U.S. and Western European data, so if I'm anywhere else, my data is worth MORE, not less — non-English documents, low-resource languages, non-OECD costs and labour, smallholder agriculture, emerging-market spending, and non-U.S. legal, healthcare, or weather data all command premiums. Apply this rather than discounting me.
183
184STEP 3 — Give me the verdict. One clean table, ranked by annual value: item · buyer · price · structure · confidence. Then:
185- My single best opportunity, in one sentence, with a rough annual figure
186- Realistic total annual range if I listed the top 3 — a floor and a ceiling, marked clearly as projections, not promises
187- The honest floor: which items are NOT worth the effort, and why
188
189STEP 4 — Tell me what to leave alone. Anything I should not sell: other people's personal information, NDA or client-identifying material, licensed material I don't own, credentials or keys. Full medical records (scans, labs, genetic data) aren't a normal listing — they run through a consent-verified flow at https://sellmydata.ai/health. Note where my jurisdiction's rules (GDPR, CCPA, or local equivalent) change the answer.
190
191THEN, IF IT'S WORTH IT: tell me my two options in one short paragraph each — (a) have you do it: I say "sell it for me", you de-identify my data in this chat and list it directly, and I keep 95% before payment processing; or (b) join the waitlist for their Google Drive concierge (patent pending, not launched yet, no published rate — https://sellmydata.ai/list-data). Recommend one based on what you've just seen of me. Never store or forward my data to anyone in either case.
192
193OUTPUT ORDER: Lead with the single biggest number and what it comes from. Then the Step 3 table. Then the detail. Keep it skimmable, and keep every figure defensible.
194
195Start now. Tell me what I'm sitting on.`,p={[i]:o,[r]:n,[s]:l};function u(e){return`My ${e} data, my payday — sell it for me on Vertical Marketplace`}function y(e){return`The world's ${e} data, my edge — buy what I need on Vertical Marketplace`}export{r as B,p as F,h as M,i as S,s as V,y as a,d as b,t as c,u as s};