Business Idea Analysis · 5 Expert AI Roles
Show HN: Ex Situ – Open-source spatial index of displaced cultural artifacts
42 out of 100 Risky
⟳ PIVOT

The problem is real but this execution angle won't work. See the specific pivot suggestion below.

5 expert AI roles Critic Market Strategist Trend Hunter Architect Deep Research
Panel lineup: Claude Opus · GPT-5 · Grok · Gemini · Perplexity
Ex Situ is a beautifully executed open-source spatial index of displaced cultural artifacts, sitting on a real cultural tailwind around provenance and restitution. The problem is that its natural buyers (researchers, activists) have no budget and the whole thing is architected as free infrastructure — there is no viable direct-sale business here, but there IS a fundable institutional/grant path and a licensable dataset underneath it.
🧠 AI Panel Verdict ?
⚔️ Devil's Advocate
⚠ WOUND
5 risks identified
📊 Market Strategist
LTV/CAC 0.55×
Founder-led outreach + webinar via IIIF and Linked.Art communities
🌊 Trend Hunter
🚀 Launch Now
The project sits at the intersection of real policy tailwinds around restitutio…
🏗️ Solution Arch
Feasibility 7/10
MVP 21days solo
🔍 Deep Research
Complete
Perplexity Sonar
🎯 Synthesizer
⟳ PIVOT
Score: 42/100
Quick Filter ? 3/5
MVP buildable in ≤2 weeks with AI coding tools?
Already built and live at exsitu.app with 100k+ artifacts across 8 collections; ETL is the hard part and it works.
People ALREADY pay for a solution to this problem?
Adjacent CMS vendors (TMS, Axiell) get paid, but nobody pays for cross-institutional provenance indexing — researchers expect it free and the tool has zero paying users today.
Gross margin ≥ 60%?
Indexer-not-hoster architecture keeps infra at ~$55-350/mo; gross margin ~86% if revenue ever materializes.
Scales without linear cost growth?
URL-pointing, no image hosting; scaling to millions needs MVT vector tiles but cost stays sublinear.
Clear competitive advantage vs free alternatives?
Wikidata, Europeana, Getty Provenance Index and Google Arts & Culture are free and better-funded; the moat is curatorial (origin-first model), which is thin and cloneable in ~3 weeks.
📋 Score Breakdown ?
Сила боли
5
Платёжеспособность ICP
3
Доступность канала
6
Юнит-экономика
2
Конкурентный ров
4
Скорость сборки
8
AI-ускорение
8
Скорость до выручки
3
Регуляторный риск
6
Тайминг тренда
7
Recommended Pivot ?
Stop treating this as a self-serve SaaS. Reposition it as (a) a grant-funded research infrastructure with an institutional anchor partner (Mellon, EU Horizon, a university digital humanities center) that buys 2-3 years of runway, and (b) a licensed 'provenance data bundle + API' sold to the ~250 high-stakes restitution/legal/auction/insurer orgs at $15k-50k/yr who actually have budget and need defensible, citable data. Keep the app free and open; monetize the curated cross-institutional dataset and reliability SLAs, not the map.
⟳ Validate this alternative idea
The analysis found a specific alternative where the blockers above don't apply. Same depth as your original report — 5 expert AI models for just $10. Available once.
⚔️ Devil's Advocate ?
No revenue model, pure passion project
High
This is an open-source, self-hosted academic tool with zero monetization path described. It's a 'labor of love' that will consume your life without ever paying rent — the classic solo-founder trap of building a product nobody will fund.
Probability:
85%
💡 Before adding a single feature, define whether this is a grant-funded nonprofit, an institutional SaaS, or a portfolio piece — and stop pretending it's a 'business' if it isn't.
Vitamin, not painkiller, for tiny niche
High
The addressable market is provenance researchers, decolonial scholars, and repatriation activists — a few thousand people globally who have no budget and expect everything free. This is intellectually important but commercially near-worthless.
Probability:
80%
💡 Interview 15 potential paying stakeholders (museum ethics departments, source-nation cultural ministries, law firms handling restitution) to see if anyone will pay for structured provenance data.
Dependency on fragile museum APIs
Medium
Your entire data layer depends on museum open-access APIs that can change, deprecate, rate-limit, or vanish overnight — and 'giving responsibility to the source institutions' means your index breaks when they update a schema.
Probability:
65%
💡 Build API health monitoring and versioned snapshots so a source change doesn't silently corrupt or empty your index.
Political controversy without institutional shield
Medium
Mapping 'displaced' artifacts as arcs from origin to Western institutions is an inherently political accusation. Museums may cut off API access, and you have no legal or institutional backing as a solo builder.
Probability:
55%
💡 Partner with an established academic institution or NGO for legitimacy and legal cover before the framing draws institutional pushback.
Solo maintenance burnout over years
High
You've been building solo since 2022 with 'a little funding.' The Next.js + Deck.gl + Strapi + Python ETL stack is a maintenance burden that will crush one person when APIs change and dependencies rot.
Probability:
70%
💡 Recruit 2-3 co-maintainers or fold the project into an existing digital humanities collective before you burn out and it dies.
Hidden Assumptions
Researchers need and will use a cross-institutional spatial index
Serious provenance researchers work case-by-case with primary archives and legal documents; a flat spatial map of API-sourced metadata is a discovery aid at best, not a research tool they'd depend on. Metadata quality from museum APIs is notoriously inconsistent and often lacks the granular provenance chains that matter for restitution.
Museum open-access APIs will remain available and cooperative
Museums control their APIs and can restrict access, especially to a project that frames their collections as 'displaced.' Your no-hosting/URL-pointing architecture makes you fully hostage to their infrastructure decisions.
Open source + AGPL + free = sustainability
Open source is a license, not a business model. AGPL actively deters commercial adoption, and 'a little funding' is not a runway. Most solo open-source cultural-heritage projects go dormant within 3 years of the founder losing steam or income.
⚠️ Cognitive Bias Check
Sesgo de costo hundido
Started as an MA thesis in 2022 and 'kept building since, mostly solo' — years invested create pressure to continue regardless of external validation.
✅ Reality check: Ask: if you started today from zero, knowing what you know, would you commit two more years solo? If no, the past investment shouldn't decide the future.
Confirmation Bias
The Show HN framing and decolonial mission attract enthusiastic agreement from a sympathetic crowd, which is easily mistaken for market demand.
✅ Reality check: Upvotes and 'this is important' comments are not usage or funding. Track weekly active researchers and any willingness-to-pay signal instead.
Optimism Bias
Framing 'prototype to production app' and indexing 100k artifacts as progress, while the fundamental sustainability and adoption questions are unaddressed.
✅ Reality check: Define concrete failure criteria (e.g., <50 monthly active users after 6 months) and commit to pivoting or archiving if hit.
🤖 AI Commoditization Risk
Days to Clone
21
Big Tech Risk
Low
The core ETL-and-map is cloneable in ~3 weeks, but the real moat is curatorial: the normalized cross-institutional data model, the origin-first taxonomy decisions, and domain trust. That's a modest but genuine moat — Big Tech has zero interest in this niche.
Worst Case
In 18 months you're still the sole maintainer, three of the eight museum APIs have changed and silently broken half the index, and the grant money ran out. The project gets a few hundred GitHub stars and admiring tweets from digital humanities Twitter, but no institution adopts it, no one pays, and you eventually archive the repo while working a day job — a beautiful, unfinished monument to a problem nobody funded.
Minimum Experiment
Email 20 named provenance researchers, museum ethics officers, and source-nation cultural ministry staff. Show them the tool and ask one question: 'Would your institution pay for or formally endorse this?' If fewer than 3 say yes within two weeks, you have confirmation this is a portfolio/grant project, not a business — and should plan accordingly.
💡 Alternative Cost
1
Package the normalized cross-museum provenance dataset and sell/license it to digital humanities departments and legal restitution teams
The data itself may have value even if the app doesn't; monetizing the dataset is a faster path to sustainability than maintaining a full app.
2
Apply for dedicated cultural-heritage grants (Mellon Foundation, EU Horizon, Digital Humanities funds) with an institutional partner
This project's true funding source is grants and institutions, not a market. One serious grant buys years of runway and legitimacy that HN upvotes never will.
3
Turn the work into a high-visibility portfolio piece and consulting practice in data-driven cultural heritage / civic tech
As a designer, this project is a stunning demonstration of skill; leveraging it into paid consulting or a design role monetizes your effort far more reliably than the tool ever will.
📊 Market & Competition ?
TAM
$0.041B
total market
SAM
$9.6M
reachable
SOM
$0.72M
your slice
Market Score
5/10
out of 10
Competitors
Company Price Revenue (est.) Strength Weakness
Wikidata + Wikimedia Commons (GLAM/IIIF ecosystem) Free (donation-supported) $0 for product; Wikimedia Foundation raises ~$180–$200M/yr in donations Massive contributor base and cross-institution coverage with open licenses and mature tooling (OpenRefine, QuickStatements). Data consistency and provenance specificity vary widely; origin–destination relationship modeling is uneven and requires expert curation.
Europeana Free (publicly funded APIs and portal) Public funding program; widely cited at ~€25–30M/yr across initiatives Aggregates tens of millions of European cultural objects with IIIF support and mature APIs. Primarily Europe-focused; not purpose-built for displaced-artifact origin→destination mapping or repatriation workflows.
Getty Provenance Index (Getty Research Institute) Free (grant-funded) N/A (grant-funded; part of Getty Trust activities) Authoritative provenance datasets and institutional credibility across museums and scholars. Scope centers on Western art market records; limited geospatial origin–destination arcs and constrained data export/licensing.
Google Arts & Culture Free N/A (no direct revenue; backed by Google) Global reach, excellent UX, and ingestion relationships with many major institutions. Not provenance- or repatriation-focused; limited research-grade export and minimal origin-site centric search.
Axiell/TMS/EMu (incumbent collection management vendors) $30k–$250k+/yr (licenses, hosting, services) Axiell Group estimated >$100M revenue across library/museum verticals Deep embed in museum workflows, procurement familiarity, SSO/SLA, and integration with collection systems. Vendor lock-in, slow feature velocity, and limited cross-institution connective tissue for origin→destination indexing.
Ideal Customer Profile (ICP)
Who
Provenance/restitution researchers and collections data leads at Western mid‑to‑large museums (≥50 FTE, ≥50k objects, using IIIF and TMS/EMu/CollectionSpace) and digital scholarship librarians at R1 universities with archaeology/art history programs.
Pain
Manual, error‑prone cross‑site provenance work using spreadsheets and ad‑hoc scripts; no origin‑first UI; weak links across museum APIs; difficulty exporting defensible, citable provenance bundles for committees, journalists, and legal teams.
Budget
Museums/universities: $5k–$20k/yr for hosted indexing, data syncs, and support; restitution/legal/advisory: $15k–$50k/yr for prioritized ingestion, bulk exports, and SLAs; individual researchers: $10–$20/mo for advanced filters/exports.
Unit Economics
ARPU
$295
/mo
LTV 12mo
$1648
12-month value
CAC paid
$3000
cost per customer
LTV/CAC
0.55×
target ≥ 3
Gross Margin
86%
gross
Monthly Churn
12%
target ≤5%
💰 Pricing Options
Researcher Pro (individual)
$14
Unlimited browsing, saved queries, batch CSV/MD exports up to 5k rows, rate-limited API key, basic email support.
~1.6% conversion
Low-friction upgrade for grad students/journalists who value export convenience but lack institutional budgets.
Organization Standard (hosted)
$249
Managed hosting, daily ETL from supported museum APIs, 250k monthly tile views, shared team workspace, webhooks, priority support (48h), permissive bulk exports for internal use.
~0.6% conversion
Hits typical small‑to‑mid museum/university procurement threshold without board approval; priced below incumbent CMS add‑ons.
Enterprise/Restitution
$1250
SSO/SLA (99.9%), custom ETL pipelines, dedicated ingestion queue, on‑prem/self‑host support for AGPL, legal-ready provenance bundles, audit trails, 2M+ monthly tile views, phone support.
~0.1% conversion
Targets high‑stakes teams (restitution/legal/auction/insurer) willing to pay for reliability, customization, and defensibility.
Best First Channel
Founder-led outreach + webinar via IIIF and Linked.Art communities
📈 Conversion: 1.4% 💰 Experiment cost: $950 ⏱ Days to first sale: 28 days
Your buyers congregate on IIIF Slack, MCN/GLAM lists, and Linked.Art; a targeted webinar with demo data for Met/V&A/SMB subsets plus 200–300 curated cold emails typically yields 2–4 qualified demos and 1 closed pilot in 3–5 weeks. Costs: webinar platform ($150), email/sequencer ($79), list research/enrichment tools ($350), design/video editing ($250), misc ($121).
📉 AI Market Dynamics (12 months)
New Competitors
+10
Price Pressure
-22%
CAC Inflation
+28%
📊 Base vs AI-Adjusted Scenario
ARPU down ~26% as new AI tools auto-aggregate museum records and offer free provenance summaries; CAC up ~29% from competitive bidding on niche keywords and sponsor slots at GLAM/MCN events; gross margin lower due to heavier inference/embedding and map/egress costs.
Metric Base AI-Adjusted
ARPU M12 $310 $230
CAC M12 $2800 $3600
Gross Margin 86% 78%
LTV/CAC 0.62× 0.39×
🔍 Deep Research ?
Competitive Intelligence

# Competitive Intelligence Report On Spatial Indexing Of Displaced Cultural Artifacts The emerging project Ex Situ, an open‑source spatial index that maps displaced cultural artifacts from their sites of origin to their current institutional holders, sits at the intersection of museum collections management, provenance research, and digital repatriation infrastructure.[1][3][8] Its emphasis on origin‑first search, cross‑institutional linkage, and a deliberately flat data model that avoids reproducing problematic taxonomies distinguishes it from conventional collections management systems, which are primarily designed for internal cataloguing and workflow rather than for global connective provenance analysis.[1][2][4] This report examines the competitive landscape around Ex Situ’s business idea by profiling the most relevant incumbent and adjacent systems, benchmarking their pricing approaches, and identifying market gaps that Ex Situ can exploit. The analysis focuses especially on collections management platforms such as TMS Collections, Axiell Collections, MuseumPlus, Vernon CMS, Argus, CatalogIt, CollectiveAccess, and CollectionSpace, together with the SummitShare digital repatriation initiative, as these represent the most direct or adjacent competition for institutions and researchers working with dispersed and contested cultural heritage.[4][5][7][8] Where possible, we quantify pricing tiers and revenue and highlight strengths and structural weaknesses, but we also explicitly acknowledge substantial data limitations, especially regarding detailed financials and user complaints, due to the constraints of available public sources.[12][13][16] ## Positioning Ex Situ Within The Cultural Heritage Technology Ecosystem Ex Situ is described in its “Show HN” announcement as an open‑source spatial index that treats museum artifacts as connecting arcs or hyperlinks between their origin sites and their institutional locations, focusing on Western and Euro‑American institutions whose collections are categorized under labels such as Islamic art, Asian and African art, ethnological collections, Middle East, and South America.[1] The project is architected as an indexer rather than a host, meaning that it aggregates and structures metadata while leaving images and primary records hosted by the source institutions; even images are retained only as URLs pointing back to the original museum repositories.[1] This connective tissue function is significant because major museum archives and APIs were not designed to “speak to each other,” and Ex Situ’s origin‑first user interface and cross‑collection mapping aim to expose the relationships between origins and destinations rather than reproduce any single institution’s cataloguing logic.[1][4] Technically, Ex Situ is built as a fully open‑source, self‑hosted stack under the AGPL‑3.0 license, using Next.js and Deck.gl on the front end, a Strapi content backend, and Python ETL pipelines that ingest data from museum open‑access APIs.[1] The index currently covers more than 100,000 artifacts across at least eight collections, including the Metropolitan Museum of Art, the Victoria and Albert Museum, and the Staatliche Museen zu Berlin, giving it a meaningful initial corpus for provenance research and spatial visualization.[1][3] A recent addition of Markdown export functionality allows researchers to download provenance data for filtered sets of artifacts, reinforcing the tool’s role as an infrastructure layer that helps researchers assemble evidence rather than as a repository that absorbs and redefines institutional records.[1] The conceptual foundations of Ex Situ align closely with contemporary provenance research in major museums such as the Staatliche Museen zu Berlin, where provenance research is explicitly framed as an inquiry into how objects were collected, acquired, and sometimes misappropriated, and as an effort to illuminate the circuitous paths objects have taken to reach their current institutional context.[3] The Berlin museums emphasize that provenance research examines questions of rightful ownership, the circumstances of acquisition, and historical injustices, and recognize that many objects in ethnological, archaeological, or art-historical collections have complex and contested histories involving colonial contexts, war, or forced sales.[3] Ex Situ’s focus on displaced artifacts, origin‑site mapping, and institutional categorization under labels like Islamic or African art suggests that it is oriented toward similar questions of misappropriation and displacement, but at a cross‑institutional, spatially aggregated scale that traditional single‑museum systems do not support.[1][3] The Ethereum Foundation’s exploration of digital repatriation also situates Ex Situ’s concept within a broader movement to use digital tools to reconnect heritage communities with artifacts that remain physically housed in Western museums.[8] In the “Digital Repatriation” post, the Foundation describes SummitShare as a system that enables museums and galleries to create exhibitions that blend physical and digital elements, with both linked via smart contracts that encode relationships and rights in the blockchain.[8] The authors frame digital repatriation as a way to provide African communities with access and agency around artifacts held abroad, including through co‑curation and shared storytelling, even when legal and logistical constraints prevent immediate physical return.[8] While SummitShare is focused on exhibitions and digital assets rather than on an index of physical displacement, its orientation toward connecting institutions, origins, and communities through digital infrastructure overlaps conceptually with Ex Situ’s goals and points to a growing ecosystem of tools that attempt to bridge archival silos. Within this broader ecosystem, Ex Situ is unusual in several ways. It is explicitly open‑source and self‑hosted under a copyleft license, in contrast to many proprietary collections management systems that require commercial licensing or subscription fees.[1][4][5] It is also architected around a flat data model, avoiding hierarchical or classificatory schemas that might embed colonial or orientalist taxonomies; instead, it indexes relationships between origin sites and destination collections and delegates responsibility for detailed object records back to source institutions.[1] This approach differs markedly from standard museum collections management systems, which are optimized for cataloguing, loans, inventory, conservation, and internal workflows, and which generally encode a rich variety of taxonomies and object types to support curatorial practice.[2][4][5] As a result, Ex Situ does not compete directly with these systems for routine collections management tasks, but it competes indirectly in the sense that institutions and researchers may allocate attention and budgets either to internal systems or to cross‑institutional analytical tools. The project’s origin as a 2022 MA thesis and its development largely by a single designer‑developer with limited funding underline that it is at an early, experimental stage rather than yet a commercial product.[1] Nonetheless, its ability to ingest museum open APIs, maintain a growing corpus of displaced artifacts, and provide exportable research data positions it as a potential backbone for more scalable provenance analytics, policy debates on repatriation, and community‑driven investigative work. In competitive terms, this means Ex Situ is not directly fighting mature enterprise vendors on feature breadth or institutional penetration; instead, it is vying for conceptual territory—namely, the notion that provenance, displacement, and spatial relationships between origins and Western museums deserve dedicated infrastructure distinct from generic collections management. ## Overview Of The Collections Management And Provenance Software Landscape The competitive field around Ex Situ is best understood as an overlapping set of segments: traditional collections management systems used by museums and cultural institutions; open‑source cataloguing platforms; digitally mediated repatriation or exhibition systems; and the broader methodological practice of provenance research within museums.[2][3][4][5][7][8] While Ex Situ does not aim to replace internal collections management, its value proposition as a spatial index of displaced artifacts intersects with several of these segments, meaning that institutions may perceive it as complementary, redundant, or alternative depending on their existing workflows and priorities. Established collections management systems such as TMS Collections by Gallery Systems, Axiell Collections by Axiell ALM Ltd, MuseumPlus by Zetcom, Vernon CMS by Vernon Systems, and Argus by Lucidea have long served as backbone software for cataloguing, managing loans, tracking conservation, and documenting exhibitions in museums worldwide.[6][7][9][14][17] Gallery Systems emphasizes that it has developed collections management solutions for the “finest cultural institutions throughout the world” for roughly forty years, indicating a deep legacy in servicing major museums with enterprise‑grade systems.[6][13] Axiell’s Collections platform is similarly positioned for museums and archives, offering modules for reporting, audit trails, movement management, public internet servers, and APIs, with detailed pricing structures published for UK public sector procurement.[12] MuseumPlus is described as supporting the complex needs of museums with a comprehensive, flexible standard application that provides real‑

Market & Risks

# Market Sizing and Risk Analysis for Open-Source Spatial Indexes of Displaced Cultural Artifacts: The Case of Ex Situ The Ex Situ project represents an emerging category of digital heritage infrastructure: an open-source, spatially explicit index that connects museum-held artifacts to their sites of origin, focusing on Islamic, Asian, African, ethnological and other non-Western collections in Euro-American institutions.[8] By design, it functions as connective tissue rather than a repository, using museum open-access APIs, a flat data model, and an origin-first search interface to let researchers traverse institutional silos and retrieve provenance-relevant information, now for more than 100,000 artifacts across eight major collections including the Metropolitan Museum of Art, the Victoria and Albert Museum, and the Staatliche Museen zu Berlin.[8] Within the broader context of growing demands for provenance transparency, repatriation debates, and interoperable cultural heritage data platforms, Ex Situ sits at the intersection of museum software, digital humanities research infrastructure, and geospatial information systems.[2][6][9][10] Using available market research on museum software, collections management tools, and the wider digital platforms sector, this report estimates that the relevant global market is modest but non-trivial, bounded by a museum-oriented software segment valued around 0.42 billion USD in 2024 and embedded in a rapidly growing digital platforms market projected to expand from 507.99 billion USD in 2026 to 1471.4 billion USD by 2035.[3][11][12] Within this space, Ex Situ’s serviceable addressable market is concentrated in Western museums with open APIs and active provenance programs, as well as university-based researchers and cultural heritage NGOs, implying a realistically capturable slice in early years that is measured more in dozens of institutional adoptions than in mass-market user counts.[2][6][7][9] A review of adjacent initiatives such as the Arches Project, the Getty Provenance Index, Rekrei, and Navigating.art reveals no directly comparable ventures that have clearly failed and shut down, but also suggests that sustainability in this domain typically depends on institutional or philanthropic backing rather than pure venture-scale returns.[9][10][13][15] Regulatory and legal risks center on data protection compliance, intellectual property and database rights attached to museum metadata and images, and the sensitivities of cultural property law and restitution debates, which are already salient in provenance research and UNESCO-led heritage mapping.[2][6][10][15] Recent funding and corporate activity, notably Lucidea’s 2024 acquisition of Eloquent Systems in the knowledge management space and continuing investment in museum and digital platform markets, indicates a cautious but real appetite for heritage- and research-oriented software, albeit one that prioritizes stable B2B and institutional models over speculative consumer plays.[3][11][12][18] Throughout, this report highlights substantial limitations: there is no dedicated market report for “spatial indexes of displaced cultural artifacts,” data on failed companies in this exact niche is sparse, and many numerical estimates must therefore be framed as scenario-based approximations rather than precise measurements. ## Context and Positioning of the Ex Situ Concept ### Conceptual Overview of Ex Situ as a Spatial Provenance Index Ex Situ was initially conceived as a Master’s thesis project in 2022 and has since been developed largely by a single designer-developer into a production-grade application, funded only modestly and released under the AGPL-3.0 as fully open-source, self-hosted software.[8]

Demand Signals

⚙️ Technical Feasibility ?
Feasibility Score
70%
Impossible Hard Easy
Days to MVP
21
solo developer
Scalability
Moderate
Client-side rendering of 100k+ arcs using Deck.gl works well, but scaling to millions of artifacts will require migrating from flat JSON endpoints to dynamic vector tiles (MVT) generated directly from PostGIS to prevent browser memory/WebGL crashes.
Recommended Stack
Next.js Deck.gl Supabase (PostgreSQL/PostGIS) Python Anthropic API
🚫 NOT in MVP ?
User accounts and saved artifact collections
💭 Standard feature for retention and personalization.
→ You are building an indexer, not a social network. Focus strictly on the data visualization and origin-first search experience.
Temporal animation slider (Timeline)
💭 Beautiful way to visualize the exact year artifacts were moved.
→ Requires highly precise acquisition dates which are frequently missing or incredibly messy (e.g., '18th century' vs 'circa 1750').
Integrating 10+ museum APIs initially
💭 Provides a massive, impressive dataset for launch.
→ Mapping just 2 complex APIs thoroughly validates the data model and UI architecture. Focus on depth over breadth for the MVP.
Key Integrations
Museum Open-Access APIs (Met, V&A, etc.)
Core data source; highly inconsistent taxonomies and schemas between institutions.
$0/mo
High
Mapbox or MapTiler
Basemaps for the Deck.gl spatial visualization.
$0/mo
Low
Anthropic API
Normalizes unstructured provenance text into clean Lat/Lng origin coordinates during the ETL process.
$30/mo
Medium
☁️ Infrastructure Cost
Stage Total/mo Breakdown
M1 (~10) $55 Supabase $25 + Anthropic API (ETL runs) $30 + Vercel free tier $0 + Mapbox free tier $0
M6 (~100) $125 Supabase $25 + Anthropic API $50 + Mapbox/MapTiler $50 (increased traffic)
M12 (~1K) $350 Supabase $100 (larger DB compute for spatial queries) + Anthropic API $100 + Mapbox $120 + Vercel Pro $30
📅 Weekly Build Plan
W1
ETL & Data Architecture
→ Python pipeline fetching, LLM-parsing, and storing 2 museum datasets into PostGIS.
~35h
W2
Frontend Spatial UI
→ Next.js + Deck.gl map rendering connections natively from the database.
~40h
W3
Search, Polish & Export
→ Responsive origin-first search menu, Markdown export out, and public deployment.
~30h
🤖 AI Build Advantage
LLMs dramatically accelerate the complex ETL pipeline by parsing messy, unstructured provenance text (e.g., 'Excavated near Luxor, 1894') into clean spatial coordinates, replacing hundreds of hours of manual curation and regex writing.
⚠️ Biggest Tech Risk
Data availability and structure. If museum APIs heavily rate-limit, change schemas without warning, or lack detailed origin descriptions, the fully automated ETL pipeline will fail to generate valid geographic coordinates for the arcs.
🛠️ MVP Build Plan ?
Days to MVP
18
solo dev
Infra Cost
$60
/month
Invest to Breakeven
$1500
P50 realistic
Tech Stack
Next.js Deck.gl Strapi Python (ETL) PostgreSQL Stripe Railway/VPS Cloudflare CDN
MVP Features
MUST
Origin-first поисковый интерфейс
Ключевая гипотеза продукта — исследователи хотят искать артефакты от места происхождения к институту, а не наоборот. Без этого UI нет отличия от каталогов музеев. Валидирует, действительно ли origin-first подход решает боль исследователей провенанса.
⏱ ~24h
MUST
Карта с дугами Deck.gl (origin → institution)
Визуализация связей — эмоциональный и репутационный крючок проекта, то, что даёт вирусность на HN/Twitter. Показывает масштаб перемещения культурного наследия одним взглядом, конвертирует посетителя в подписчика.
⏱ ~20h
MUST
Python ETL из open-access API музеев
Без данных нет продукта. Индексация 3-8 коллекций через открытые API — это фундамент. Критично проверить, что нормализация к плоской модели реально масштабируется на разные схемы (Met, V&A, SMB).
⏱ ~40h
MUST
Экспорт провенанс-данных (md/CSV) по фильтру
Единственная фича с прямой платёжной готовностью: исследователи и институции нуждаются в структурированной выгрузке для публикаций и due diligence. Это то, за что академия и юристы реституции реально заплатят.
⏱ ~12h
MUST
Фильтры по региону происхождения / категории / институту
Плоская модель бесполезна без навигации. Фильтры превращают 100k записей в исследовательский инструмент и напрямую питают экспорт. Валидирует, какие срезы данных наиболее востребованы.
⏱ ~16h
MUST
Ссылки на источник + кэш метаданных (без хостинга изображений)
Юридически и этически критично: индексатор, а не хостер. Снижает правовые риски (AGPL, авторские права музеев) и стоимость инфраструктуры. Изображения — только URL на институт.
⏱ ~8h
SHOULD
Регистрация + план подписки для институций/исследователей
Нужен, чтобы отделить бесплатный просмотр от платного экспорта/API. Без стены монетизации невозможно проверить готовность платить. Простой Stripe + email-auth достаточно для MVP.
⏱ ~14h
🗺️ First Customer Journey ?
1
Обнаружение
👤 Видит Show HN пост или расшаренную карту в Twitter/академической рассылке
👁 Заголовок 'карта перемещённых культурных артефактов' + скриншот дуг на карте мира ⚙️ Запуск на HN, шеринг визуализации, аутрич к digital humanities сообществам
2
Первое взаимодействие с картой
👤 Открывает карту, исследует дуги, ищет свой регион/страну происхождения
👁 Интерактивная визуализация, origin-first поиск, масштаб проблемы ⚙️ Быстрая загрузка Deck.gl, понятный onboarding без регистрации
3
Осознание ценности исследования
👤 Фильтрует по региону/категории, кликает на артефакт, переходит к источнику
👁 Метаданные провенанса, ссылка на институт, релевантные срезы данных ⚙️ Качество нормализации данных, точность связей origin→institution
4
Попытка экспорта ⚠️ DROP RISK
👤 Хочет выгрузить отфильтрованный набор провенанс-данных для исследования
👁 Кнопка экспорта → стена регистрации/подписки для больших наборов ⚙️ Разделение бесплатного просмотра и платного экспорта, Stripe checkout
5
Оплата и получение ценности
👤 Оформляет подписку (или институция покупает лицензию), скачивает md/CSV
👁 Структурированный экспорт готовый для публикации/анализа ⚙️ Надёжная генерация файлов, квитанция, институциональный биллинг
6
Удержание
👤 Возвращается за новыми коллекциями, подписывается на обновления индекса
👁 Новые проиндексированные институции, email-дайджест изменений провенанса ⚙️ Регулярное добавление коллекций, changelog, уведомления
💡 Dropout mitigation: Академическая аудитория ожидает, что open-source и данные бесплатны — стена оплаты на экспорте вызовет массовый отток. Решение: сделать базовый экспорт (до N записей, watermark/атрибуция) бесплатным, а платить должны только за объёмные/коммерческие/API выгрузки и институциональные лицензии. Монетизацию строить на институциях (музеи, университетские библиотеки, юрфирмы по реституции) и грантах/пожертвованиях, а не на отдельных исследователях. Явно коммуницировать 'freemium для науки, лицензия для институций', чтобы не разрушить доверие open-source сообщества.
💰 Financial Sketch (Realistic) ?
Investment Needed
$6000
until breakeven
Breakeven
М9
month of payback
MRR М12
$3200
at month 12
LTV/CAC
0.55×
target ≥ 3
Unit Economics — Margin per Sale ?
Price per unit
$1250
Cost per unit (COGS)
$175
Platform fee
0%
Margin per unit
$1075.0
Min. price to break even: $175.0
Enterprise/restitution plan at $1,250/mo carries ~86% margin (COGS = custom ETL + support); economics are healthy per-unit but the fragility is volume — closing even 3 such accounts is the whole challenge, and self-serve $14/mo tiers won't cover CAC.
Month MRR
M1 $0
M3 $0
M6 $800
M12 ✅ Breakeven $3200
🟥 burning cash · 🟩 cash positive · ✅ BREAKEVEN = investment fully recovered
📈 Three Scenarios (P20 / P50 / P80) ?
P20 — Осторожный
MRR М12
$600
CAC
$180
Churn/mo
18%
To Breakeven
$3500
Академический рынок узкий и медленный, гранты не приходят, CAC 2× из-за длинных циклов институций. Органика с HN даёт всплеск трафика, но почти нулевую конверсию в платящих. Экспорт остаётся бесплатным для большинства.
P50 — Реалист
MRR М12
$2800
CAC
$60
Churn/mo
8%
To Breakeven
$1500
Смешанная модель: пожертвования/OpenCollective (проект open-source) + платные институциональные лицензии на API/экспорт ($50-200/мес) + гранты цифровых гуманитарных наук. CAC низкий за счёт органики HN/Twitter и академических рассылок. Owned-asset: список email из HN-запуска + существующая аудитория проекта с 2022 (контент ~$100/мес на поддержку).
P80 — Оптимист
MRR М12
$8000
CAC
$15
Churn/mo
4%
To Breakeven
$500
HN-запуск попадает в топ, тема реституции резонирует в прессе (Guardian, Hyperallergic). Крупный грант ($20-50k) или партнёрство с университетской библиотекой. Owned-asset: вирусная карта как органический канал (шеринг в соцсетях, стоимость ~$50/мес хостинг/контент). Институции подписываются на командные лицензии.
Month P20 P50 realistic P80
M1 $0 $50 $200
M3 $80 $300 $900
M6 $250 $900 $3000
M12 $600 $2800 $8000
🧪 Hypotheses to Validate ?
H1
If we approach restitution/legal/auction/insurer teams, at least 3 of 20 will confirm willingness to pay $15k-50k/yr for defensible, exportable provenance bundles with an SLA.
🔬 20 targeted outreach emails/calls to named restitution lawyers, auction-house provenance leads, and cultural ministries showing a legal-ready export demo. ⏱ 21 days
H2
If we pitch this as open research infrastructure with an institutional partner, at least one grant program or university center will express concrete interest in co-applying.
🔬 Reach out to 5 digital humanities centers + shortlist 3 grant programs (Mellon, EU Horizon, NEH); pitch a co-application. ⏱ 30 days
H3
If APIs are monitored, the index can stay >90% valid across the 8 current collections for 60 days without silent breakage — proving the data layer is reliable enough to sell an SLA on.
🔬 Add API health monitoring + versioned snapshots; log breakage/drift over 60 days. ⏱ 60 days
🛑 Kill Criteria ?
Fewer than 2 of 20 high-stakes (legal/restitution/auction/insurer) contacts express concrete willingness to pay within 6 weeks of outreach.
No grant program or institutional partner agrees to a co-application or letter of intent within 90 days.
Two or more of the 8 source museum APIs restrict access or break schema in a 60-day window, confirming the data layer is too fragile to sell reliability on.
⚖️ Risks & Opportunities ?
Top Risks
No paying customer exists today — the ICP that needs it (researchers/activists) has no budget, and the ICP that has budget (legal/restitution) hasn't been validated or approached.
Full dependency on fragile museum APIs that can rate-limit, change schema, or cut access — especially given the politically charged 'displaced artifacts' framing.
Solo-maintainer burnout across a Next.js + Deck.gl + Strapi + Python stack that has already run since 2022 on 'a little funding', with no runway plan.
Top Opportunities
Restitution/legal/auction/insurer segment (~250 orgs, $15k-50k/yr) needs defensible, citable, exportable provenance bundles — a real willingness-to-pay niche the current app ignores.
Cultural-heritage grants (Mellon, EU Horizon, digital humanities funds) reward exactly this kind of open infrastructure with an institutional partner — years of runway, not upvotes.
The normalized cross-institutional dataset itself is licensable to digital humanities departments even if the app never monetizes.
Next 48 Hours ?
1
Build a target list of 20 named contacts across restitution law firms, auction-house provenance departments, source-nation cultural ministries, and insurers — with real names and emails, not generic inboxes.
2
Record a 3-minute Loom demo of the legal-ready Markdown/provenance-bundle export for a real contested-artifact example (e.g., a Benin Bronze), framed around 'defensible, citable evidence'.
3
Draft and send the first 5 outreach emails asking one question: 'Would your team pay for or formally endorse structured, exportable provenance data?'
📅 30-Day Action Plan ?
W1
Week 1
Validate willingness-to-pay in the ONLY segment with budget before building anything new.
Send all 20 outreach emails to the high-stakes segment; aim for 5 booked demo calls.
On each call, ask directly about budget, procurement process, and what 'defensible provenance data' must contain to be legally usable.
Log every response in a simple sheet: interested / not / why — and count concrete willingness-to-pay signals against the H1 threshold of 3.
W2
Week 2
Test the grant/institutional path in parallel.
Contact 5 digital humanities centers and 3 grant programs (Mellon, EU Horizon, NEH) pitching co-applied open research infrastructure.
Turn the strongest demo interest into one written pilot proposal or letter-of-intent request.
W3
Week 3
Harden the data layer to make reliability sellable.
Add API health monitoring and versioned snapshots so source changes don't silently break the index.
Package the 'provenance bundle' export into a polished, citable, legal-ready format based on Week 1 feedback.
W4
Week 4
Decide: fund it as infrastructure, license the data, or archive gracefully.
Tally H1/H2 results against kill criteria; if ≥3 paying signals OR a grant LoI, commit to the institutional/dataset model and draft pricing.
If both fail, reposition Ex Situ as a portfolio/consulting showcase and recruit 2-3 co-maintainers to keep it alive without founder burnout.
⟳ Want to validate the alternative direction?
The analysis found a specific alternative where the blockers above don't apply. Same depth as your original report — 5 expert AI models for just $10. Available once.