Skip to content

PropPie Analytix (B2B RealtyAnalytix) — Product Specification

"The Bloomberg Terminal for Indian commercial real estate, where every number is one click from the underlying RERA filing or IGR deed."

1. Product identity

Field Value
Internal name PropPie Analytix
Market name RealtyAnalytix (or PropPie Analytix — OPEN QUESTION: final brand TBD)
Tagline Source-cited real-estate intelligence. Every number, one click from the document.
Status In design (foundation phase)
Product type B2B SaaS dashboard + API
Geography Maharashtra-only at launch; Karnataka pilot Q4 2027
Compliance posture Information-only; no personalised advice; no SEBI IA/RA registration required

Target users

Segment Example firms What they buy today Why they'd switch
Mid/large developers Godrej Properties, Lodha, Panchshil, Kolte-Patil, Shapoorji PropEquity, in-house research, broker whispers Source-cited competitive intelligence, delay forensics, GR tracking
Brokers / IPC firms Anarock, Knight Frank, JLL, CBRE, Cushman & Wakefield CRE Matrix, manual IGR pulls, quarterly reports Fast comparables, white-label widgets, real-time velocity
Institutional investors Embassy REIT, Mindspace REIT, Brookfield, Blackstone, AIFs, family offices Knight Frank custom reports, CRE Matrix, in-house teams Underwriting-grade data trail, probabilistic projections, audit-ready citations
Lenders SBI, HDFC, Bajaj Finance, PNB Housing, Piramal In-house appraisals, manual registry pulls Automated collateral checks, zone risk index, developer trust scores

Competitive positioning

PropPie Analytix is not a better spreadsheet. It is a source-cited, AI-native intelligence layer that treats every number as a claim requiring evidence. See competitive-landscape.md for the full map. Section 4 below covers differentiators in detail.


2. Jobs to be done

Written as queries the dashboard answers. Grouped by user segment.

Developers — competitive intelligence

  • "What is the 90-day transaction velocity for 2BHK residential in Hinjewadi, and how does it compare to Wakad and Kharadi?"
  • "Which of my competitors delivered on time in Pune in the last 3 years, and what was my percentile rank?"
  • "Which Government Resolutions in the last 6 months affect FSI or TDR in MMR, and what's their projected impact on land cost?"
  • "What price per sqft are buyers actually paying in Baner (registered deeds, not listings), and how has it moved quarter-over-quarter?"

Brokers / IPC firms — fast comparables

  • "Give me 10 comparable registered sale transactions for a 1,200 sqft office in Lower Parel, last 12 months, with source deeds."
  • "What's the median cap rate for Grade A office space in BKC vs. Andheri East vs. Airoli?"
  • "Show me micromarkets where residential velocity is up >20% QoQ — I need to brief my sales team."
  • "Generate a white-label market snapshot for Pune West that I can send to my client."

Institutional investors — due diligence at scale

  • "What's the zone risk index for Bhiwandi warehousing, and what are the title clarity scores for all projects within 5km of NH-3?"
  • "Show me the developer trust score for Promoter X, with all inputs visible, for our underwriting committee."
  • "Run a Monte Carlo projection for a 50,000 sqft warehouse in Chakan: IRR distribution at year 5 and year 10."
  • "Which micromarkets have cap rates above 8% with zone risk below 4/10?"

Lenders — project financing evaluation

  • "What is the cost overrun history and complaint density for Developer Y across all their MahaRERA-registered projects?"
  • "Show me the escrow balance proxy and sales velocity for Project Z — is there liquidity risk?"
  • "What is the title clarity score and encumbrance status for the collateral property at Survey No. 45/2, Taluka Mulshi?"

3. Core dashboard modules

3.1 Market Momentum

What it shows: Real-time pulse of the market, powered by registered transactions (not listings).

Component Description Key attributes
Velocity heatmap H3 hex-grid map coloured by mkt.transaction_velocity_90d; toggle residential/commercial mkt.transaction_velocity_90d, mkt.transaction_velocity_180d, loc.lat, loc.lng
Bullish % gauge % of micromarkets with positive QoQ velocity change mkt.sector_momentum_pct
Sector breakdown Stacked bar: residential, office, retail, warehousing, data centres proj.type, mkt.sector_momentum_pct
Price trend sparklines Per-micromarket price/sqft trend (90d rolling) mkt.median_price_per_sqft_90d, mkt.price_appreciation_yoy_pct
Lifecycle stage tags Each micromarket tagged Emerging / Growing / Mature / Cooling mkt.micromarket_lifecycle_stage

Source citation: Every velocity number links to the underlying IGR transaction list. Every price figure shows n (sample size) and date range.

3.2 Delay Forensics

What it shows: Delivery delay analysis across developers, projects, and micromarkets.

Component Description Key attributes
Delay distribution Histogram of (actual - promised) months across all projects in filter proj.promised_completion_dates, proj.actual_completion_date
Developer percentile table Rank table: developer median delay vs. comparable promoters (same city tier, project count) dev.trust_score sub-component
Root-cause tags NLP-classified reasons (regulatory, financial stress, litigation, weather, COVID) extracted from Form C progress reports and news legal.complaints_maharera, news entity tags
Cost overrun correlation Scatter: delay months vs. cost overrun % fin.cost_overrun_pct
Project-level drill-down Timeline view: original promise, each revision, actual delivery proj.promised_completion_dates, proj.actual_completion_date, proj.estimated_cost_history

Defamation guardrails: Developer names shown with percentile framing, never absolute labels. "Request a review" link on every developer card. See compliance-and-legal.md.

3.3 Yield and Cap Rate Heatmaps

What it shows: Where the money is — gross yield and cap rate by micromarket and sector.

Component Description Key attributes
Cap rate heatmap Choropleth map: median cap rate by micromarket, filterable by sector mkt.cap_rate_median
Yield benchmark table Tabular view: micromarket, sector, median yield, n, range, trend mkt.yield_benchmark
Yield vs. risk scatter X = zone risk index, Y = cap rate; bubble size = transaction count risk.zone_risk_index, mkt.cap_rate_median
Rent escalation tracker Median registered lease escalation % by micromarket lease.escalation_pct
ASR gap overlay Toggle to show ASR gap % alongside yield — spots under/over-valued zones fin.asr_gap_pct

Confidence display: Every cap rate shows sample size. Micromarkets with n < 10 transactions in 12 months are greyed out with "low confidence" label.

3.4 Comparable Analysis

What it shows: Auto-generated comparable sets for any asset, refreshing on every new IGR registration.

Component Description Key attributes
Comparable finder Input: address or RERA number. Output: 5-10 nearest-neighbour transactions ai.comparable_set
Side-by-side cards Key fields (price/sqft, area, date, developer, sector) for easy comparison fin.price_per_sqft_carpet, area.carpet_sqft, fin.sale_date
Source deed links Each comparable links to the registered deed (IGR) fin.sale_price source lineage
Auto-refresh alerts Notification when a new transaction matches a saved comparable search IGR daily feed
Export PDF / Excel export with full citations for client reports All displayed attributes

3.5 Developer Intelligence

What it shows: Comprehensive, explainable developer profiles.

Component Description Key attributes
Trust Score card 0-100 score with all five input penalties visible and expandable dev.trust_score
Portfolio view Map + list of all MahaRERA-registered projects by promoter PAN party.promoter_pan, proj.maharera_number
Delivery track record Timeline of all projects: promised vs. actual, with trend line proj.promised_completion_dates, proj.actual_completion_date
Complaint history Count, category, resolution status legal.complaints_maharera
Financial health signals Cost overruns, loan disclosures, escrow proxy fin.cost_overrun_pct, fin.loan_disclosure, fin.escrow_balance_proxy
Sentiment pulse 30-day aggregated news/social sentiment ai.sentiment_score
Plain-English summary LLM-generated developer narrative (defamation-guarded) ai.developer_track_record_summary

Compliance: Percentile framing only. Right-of-reply mechanism documented. See derived-attributes-spec.md section 12 for the math and guardrails.

3.6 Policy Intelligence Feed

What it shows: Government Resolutions classified, scored, and linked to affected micromarkets.

Component Description Key attributes
Live feed Chronological list of RE-relevant GRs with NLP classification policy.applicable_grs
Impact scoring High / Medium / Low impact badge per GR GR NLP classifier output
Tailwind / headwind flags Per-micromarket positive and negative policy signals policy.tailwind_flags, policy.headwind_flags
Affected-area map GR impact overlaid on micromarket map GR entity extraction + geospatial
Plain-English summary LLM-generated "what this means for property values in X" ai.gr_impact_summary
Alert subscriptions Users subscribe to micromarkets or departments; get notified on new GRs Notification system

Source: Every GR links to the original PDF on the government portal. Translation from Marathi provided alongside.

3.7 Material Pulse Tracker (Phase 2)

What it shows: Construction input cost trends affecting project economics.

Component Description Data source
Steel price index Monthly trend, regional SAIL / TMT bar spot prices
Cement price index Monthly trend, regional UltraTech / ACC / regional
Labour cost index Quarterly, by trade and region Industry surveys, CMIE
Sand / aggregate Regional, quarterly Local market surveys
Cost-push impact calculator "If steel rises 10%, estimated impact on per-sqft cost for a typical 2BHK project in Pune" Derived from input-cost model

Status: Phase 2 module. Data sources not yet contracted. Marked as OPEN QUESTION: build vs. license from CMIE or similar.

3.8 Risk Radar

What it shows: Multi-layered risk view per micromarket and per project.

Component Description Key attributes
Zone Risk Index map Choropleth: 0-10 risk score per micromarket risk.zone_risk_index
Risk decomposition Per-zone breakdown: legal, market, policy, environmental components Zone risk sub-scores
Title clarity overlay Toggle to show average title clarity score per micromarket legal.title_clarity_score
Environmental flags Flood zone, CRZ, forest zone, AQI overlays risk.flood_zone_flag, risk.crz_flag, risk.forest_zone_flag, risk.aqi_annual_avg
Litigation hotspots Heatmap of active real-estate litigation density legal.litigation_records aggregated
Risk alerts Notification when zone risk index changes by > 0.5 points Derived trigger

3.9 Asset Underwriting Module (Institutional)

What it shows: Single-asset deep-dive for underwriting committees. Designed to produce audit-ready output.

Component Description Key attributes
Asset profile All 140 attributes for one project/asset, organised by category Full attribute set
Title chain walkthrough Plain-English title history with source documents ai.title_chain_explanation, legal.chain_of_title
Wealth trajectory simulation Monte Carlo paths at p20/p50/p80 for 5 and 10 year horizons ai.wealth_trajectory_paths, ai.projected_irr_p50_p20_p80
Comparable basket Auto-generated comparables with full citations ai.comparable_set
Risk scorecard Title clarity + zone risk + developer trust + tenant quality in one view All risk-related derived attributes
Brutal honesty flags Red flags and "consider before investing" notes ai.brutal_honesty_flags
Export: underwriting memo PDF export formatted for IC presentation; every number cited All displayed attributes

4. Key differentiators vs. competitors

Dimension PropPie Analytix PropEquity CRE Matrix Liases Foras IPC research (JLL, CBRE, etc.)
Source citation Every number one click from source deed/filing Opaque aggregation; users can't verify Limited source visibility Report-level citations only Report-level; often proprietary surveys
Data freshness Daily IGR feed for priority micromarkets Monthly/quarterly bulk updates Monthly Quarterly reports Quarterly reports
AI-native UX Conversational queries, simulations, what-if scenarios Form-based filters + Excel export Dashboard + export Static PDF reports Custom consulting deliverables
Policy intelligence Real-time NLP-classified GR feed with impact scoring No GR tracking No GR tracking Occasional policy notes Manual analyst coverage
Delay forensics Automated, percentile-framed, root-cause tagged Basic delivery status Not available Not available Occasional case studies
Probabilistic projections Monte Carlo with visible assumptions and confidence bands Point estimates Not available Proprietary forecasts (opaque) Analyst opinion
Compliance posture Info-only by design; defamation-guarded scores Not explicitly positioned Not explicitly positioned Research product Research + consulting
Onboarding time Self-serve; productive in hours Weeks (sales cycle + training) Weeks Weeks Months (custom engagement)
Pricing Transparent tiers; accessible to mid-size firms Mid-six-figures; enterprise only Mid-five to mid-six-figures Mid-five-figures Project-based; premium
Geographic depth Maharashtra-deep (then expand) Pan-India (shallower per state) MMR + Bangalore focus Pan-India residential Pan-India

The one-sentence wedge: PropEquity has 15 years of data breadth; we have Maharashtra-depth, source-citation discipline, and an AI-native UX that makes their product feel like a filing cabinet.


5. Data dependencies — which attributes power each module

Module Critical attributes Source layers
Market Momentum mkt.transaction_velocity_*, mkt.sector_momentum_pct, mkt.median_price_per_sqft_90d, mkt.price_appreciation_yoy_pct, mkt.micromarket_lifecycle_stage, loc.micromarket_id IGR (daily), MahaRERA (monthly)
Delay Forensics proj.promised_completion_dates, proj.actual_completion_date, proj.estimated_cost_history, legal.complaints_maharera, dev.trust_score MahaRERA (quarterly + monthly)
Yield and Cap Rate mkt.cap_rate_median, mkt.yield_benchmark, lease.rent_amount_monthly, lease.escalation_pct, fin.sale_price, fin.asr_gap_pct IGR L&L + Sale (daily), ASR (annual)
Comparable Analysis ai.comparable_set, fin.price_per_sqft_carpet, area.carpet_sqft, fin.sale_date, loc.micromarket_id IGR (daily)
Developer Intelligence dev.trust_score, party.promoter_pan, legal.complaints_maharera, fin.cost_overrun_pct, ai.sentiment_score, ai.developer_track_record_summary MahaRERA (monthly), News (hourly), Social (hourly)
Policy Intelligence policy.applicable_grs, policy.tailwind_flags, policy.headwind_flags, ai.gr_impact_summary, infra.upcoming_projects GR portals (daily)
Material Pulse External cost indices CMIE, industry feeds (Phase 2)
Risk Radar risk.zone_risk_index, legal.title_clarity_score, risk.flood_zone_flag, risk.crz_flag, risk.forest_zone_flag MRSAC, MoEFCC, MahaRERA, IGR
Asset Underwriting Full ~140 attribute set + all derived scores + Monte Carlo All sources

Full attribute catalogue: data-attributes.md. Derivation math: derived-attributes-spec.md.


6. White-label and embed capability

Embeddable widgets

IPC firms and large brokerages want to embed PropPie intelligence into their own platforms and client reports. Planned widgets:

Widget Use case Embed format
Micromarket snapshot Broker sends a "market update" email/PDF to clients iframe / PDF export
Comparable card Side-by-side comparable view for a specific property iframe / API
Velocity gauge Real-time velocity indicator for a micromarket iframe / JS component
Developer trust card Developer profile summary (defamation-guarded) iframe / API
Cap rate heatmap Interactive map for a region iframe

White-label terms

  • Partner's branding on the widget chrome; "Powered by PropPie" footnote (non-negotiable for citation integrity)
  • Data citation links always visible — partner cannot strip source links
  • Pricing: per-widget monthly fee or bundled with enterprise tier

OPEN QUESTION: White-label pricing model — per-embed, per-view, or flat monthly? Requires design partner feedback.


7. API access tier

API design principles

  • RESTful, versioned (/v1/), JSON responses
  • Every response includes source, source_url, confidence, extracted_at per attribute
  • Rate-limited by tier
  • Authentication: API key + OAuth2 for enterprise SSO

Planned endpoints (indicative)

Endpoint group Example Tier
/micromarkets List micromarkets, velocity, prices, lifecycle stage Standard
/transactions Search registered transactions by filters Standard
/projects MahaRERA project data by RERA number or filters Standard
/developers Developer trust score, portfolio, track record Standard
/comparables Auto-generated comparable set for an address Premium
/risk Zone risk index, title clarity, environmental flags Premium
/policy GR feed, tailwind/headwind flags Standard
/simulate Monte Carlo projection for an asset Enterprise
/underwriting Full asset profile with all 140 attributes Enterprise

Rate limits (draft)

Tier Requests/day Concurrent Simulation
Standard 5,000 10 Not included
Premium 25,000 50 100/day
Enterprise 100,000+ Custom Unlimited

OPEN QUESTION: API-only pricing vs. bundled with dashboard access. Some B2B customers (lenders, fintechs) want API without the dashboard.


8. Pricing model options

Status: OPEN QUESTION — to be validated with design partners (Q4 2026).

Option A: Seat-based SaaS

Tier Seats Modules Price range (annual)
Starter Up to 3 Market Momentum, Policy Feed, Comparables 3-5L
Professional Up to 10 All except Underwriting 8-15L
Enterprise Unlimited All modules + API + white-label 20-50L

Option B: AUM-based (institutional investors)

  • Base fee + basis points on assets under analysis
  • Aligns cost with value for large portfolios
  • Risk: hard to verify AUM; perverse incentive to understate

Option C: Tiered by data consumption

  • Per-query or per-report pricing with monthly minimums
  • Low barrier to entry; scales with usage
  • Risk: unpredictable revenue; usage anxiety for customers

Recommendation (pre-validation)

Start with Option A (seat-based) for simplicity and predictability. Offer enterprise custom pricing for institutional investors who need API + white-label. Revisit after 10 paying customers.

Decision needed by: Q4 2026, informed by design partner conversations.


9. Go-to-market

Phase 1: Design partners (Q4 2026)

Target: 5 design partners using the product for free in exchange for weekly feedback sessions.

Partner type Target count Selection criteria What we learn
Mid-size Pune developer 2 Active, 10+ RERA projects, tech-forward Module priority, daily workflow fit
IPC firm (regional office) 1 Pune or MMR office of Anarock / Knight Frank / JLL White-label demand, comparable workflow
AIF / family office 1 Active in Maharashtra CRE Underwriting module requirements
NBFC / HFC 1 Evaluating Maharashtra project finance Lender workflow, collateral check needs

Phase 2: Paid launch (Q2 2027)

  • Convert design partners to paid
  • Outbound to 50 prospects (developer analytics heads, IPC research teams, institutional investment teams)
  • Target: 20 paying customers by end of Q2 2027

Sales approach

  • No inside sales team initially. Founder-led sales (Aishvarya for product, Vishal for technical credibility).
  • Demo-first: 30-minute live demo on prospect's own micromarket data. "Here's what we know about your last 3 projects — did you know this?"
  • Free trial: 14-day full access with onboarding call.
  • Annual contracts with quarterly payment option.

Channel partnerships (Phase 2+)

  • IPC firms as distribution channel: they recommend PropPie to their developer/investor clients
  • Lending platforms: embedded risk scoring via API
  • CA/legal firms: referral for due-diligence workflows

OPEN QUESTION: Commission/referral structure for channel partners. See open questions Q31.


10. Success metrics

Metrics that matter for Analytix, aligned with vision.md.

Metric Target (12 months post-launch) Why it matters
Paying B2B customers 20+ Revenue validation
% users with daily active usage > 40% Real workflow integration, not shelf-ware
Source-citation rate 100% of factual claims Non-negotiable trust metric
Time-to-first-value (onboarding) < 2 hours from signup to first useful query Self-serve validation
NPS among design partners > 50 Product-market fit signal
API adoption (among enterprise tier) > 50% of enterprise customers using API Platform stickiness
White-label embed count > 10 active embeds Distribution channel validation
Hallucination incident rate 0 reported incidents Trust metric
Data freshness SLA compliance > 95% of attributes within stated cadence Reliability
Monthly recurring revenue 1-2 Cr ARR by month 12 Business viability

11. Phasing — MVP vs. future

Phase 1: MVP (Q4 2026 - Q1 2027)

Ship to design partners. Minimum viable intelligence.

Module Scope Dependencies
Market Momentum Velocity heatmap + price sparklines for 20 priority micromarkets (Pune + MMR) IGR daily feed operational
Delay Forensics Developer delay table + project drill-down MahaRERA full ingestion
Comparable Analysis Basic comparable finder (address input, 5-10 results) IGR + entity resolution
Developer Intelligence Trust Score + portfolio view (no sentiment yet) MahaRERA + derived pipeline
Policy Intelligence Feed GR feed with manual classification (NLP classifier v0) GR scraping pipeline
Export PDF/Excel export with citations All above

Not in Phase 1: Yield heatmaps (need L&L data depth), Risk Radar (need GIS layers), Asset Underwriting (need Monte Carlo), Material Pulse (need external feeds), API (need stable schema), white-label.

Phase 2: Full product (Q2 2027 - Q4 2027)

Module Scope Dependencies
Yield and Cap Rate Heatmaps Full cap rate and yield maps; requires L&L transaction depth Sufficient L&L volume in IGR feed
Risk Radar Zone risk index + environmental overlays MRSAC GIS layers acquired
Asset Underwriting Full 140-attribute profile + Monte Carlo + underwriting memo export All derived attributes operational
API v1 Standard + Premium endpoints Stable schema, rate limiting
White-label widgets Embeddable micromarket snapshot + comparable card Widget framework built
Conversational layer Natural-language query interface for the dashboard RAG pipeline operational

Phase 3: Scale (2028+)

Module Scope
Material Pulse Tracker Steel, cement, labour, sand indices
Karnataka expansion Bangalore micromarket data foundation
Alert engine Custom alerts on velocity, price, GR, developer changes
Benchmarking reports Automated monthly/quarterly market reports per micromarket
Lender integration module Collateral evaluation API for lending platforms

12. Risks and mitigations

# Risk Likelihood Severity Mitigation
A1 IGR data feed insufficient — bulk licensed feed too expensive or too stale Medium High Start with self-scrape for priority micromarkets; license selectively; budget for 5-25L/year
A2 PropEquity undercuts on price after we launch Medium Medium We compete on depth + UX + citation, not price; their architecture can't easily add citation
A3 Design partners don't convert to paid Medium Medium Ensure design partners represent real buyer personas; build for their workflow, not ours
A4 Developer defamation risk from Trust Score module Low-Medium High Percentile framing, right-of-reply, legal review of display language. See compliance-and-legal.md
A5 GR classifier accuracy too low for production Medium Medium Start with human-in-loop classification; automate incrementally; monitor precision/recall
A6 Enterprise sales cycle too long (6-12 months) High Medium Target mid-size firms first (shorter cycle); offer monthly plans; founder-led sales
A7 Data depth insufficient in non-priority micromarkets Medium Medium Be transparent about coverage gaps; show confidence levels; expand micromarket priority list based on demand
A8 Hallucination in LLM-generated summaries Medium High Eval framework, source-citation mandate, "I don't know" defaults, human review for launch
A9 SEBI regulatory change pulls dashboard intelligence into RA territory Low High Info-only by design; no "buy/sell" signals; legal opinion on file. See risks-and-open-questions.md R1
A10 Team bandwidth — 5 developers split across 3 products High Medium Sequence strictly; Analytix MVP uses same data pipeline as Broker; shared infrastructure

13. Cross-references

Document Relevance
SOUL.md Honest broker principles apply to all B2B outputs
vision.md Strategic milestones: design partners Q4 2026, paid launch Q2 2027
competitive-landscape.md Lane C — full competitor analysis for B2B analytics
moat-and-differentiation.md Five layers of moat; Analytix leverages all five
compliance-and-legal.md SEBI info-only stance, defamation guardrails, DPDP
risks-and-open-questions.md Q6 (design partner segment), Q8 (pricing model), Q31 (IPC partnerships)
data-attributes.md Full ~140 attribute catalogue powering all modules
derived-attributes-spec.md Math specs for every derived score
data-sources.md Source catalogue, access patterns, costs, gotchas
data-quality-framework.md Confidence, lineage, freshness SLAs
pipeline-spec-for-vishal.md Pipeline contract — what Analytix needs from the pipeline

Open questions summary

All open questions from this document, collected for tracking.

# Question Owner Needed by
OQ-A1 Final market-facing brand: "PropPie Analytix" vs. "RealtyAnalytix" vs. other Aishvarya Q3 2026
OQ-A2 Pricing model: seat-based vs. AUM-based vs. consumption-based Aishvarya + design partners Q4 2026
OQ-A3 White-label pricing: per-embed, per-view, or flat monthly Aishvarya + IPC design partner Q4 2026
OQ-A4 API-only tier: separate pricing vs. bundled with dashboard Aishvarya Q1 2027
OQ-A5 Material Pulse data source: build vs. license from CMIE or similar Vishal Q2 2027
OQ-A6 Design partner selection: which specific firms to approach first Aishvarya Q3 2026
OQ-A7 Channel partner commission/referral structure Aishvarya Q2 2027
OQ-A8 IGR bulk feed: self-scrape vs. PropEquity/CRE Matrix license vs. hybrid Vishal Q3 2026

These should be mirrored in risks-and-open-questions.md and resolved per the cadence defined there.


This spec is the source of truth for PropPie Analytix product decisions. Implementation details (tech stack, database, orchestration) are Vishal's domain per AGENTS.md section 5.3. This document defines the what and why; the pipeline spec defines the contract; Vishal defines the how.

Last updated: foundation session, May 2026.