Private Beta — Limited Early Access Available

Ground-Truth ABS Data.
Rigorous by Design.
AI-Native by Default.

Built by practitioners with 35+ years across top-tier investment banks and leading hedge funds — because we spent decades on the other side of that invoice. 100M+ agency loans, MISMO-aligned, AI-native, and in your environment in days. Not weeks. Not a procurement cycle. Days.

100M+
Agency Loans
10+ yrs
Deep History
Days
To Integration
5–10x
Less Than Legacy

Designed by veterans of top-tier investment banks and leading hedge funds — built for the institutional credit workflows they know from the inside.

Request Early Access to TrueABS

Be among the first 50 users. Early access includes 3 months free on any paid tier.

No credit card required · Limited spots available

You're on the list!

We'll reach out within 2 business days with next steps.
TrueABS covers Fannie Mae Freddie Mac Ginnie Mae SEC EDGAR MISMO 3.x Aligned Cloud-Native Delivery

The Incumbents Work for the
Top 20 Firms. What About Everyone Else?

Three data providers control structured credit. After a decade of consolidation, pricing has moved in one direction — up. Independent analysts and boutique firms pay enterprise rates for data they can't query without a team of ETL engineers. TrueABS was built for everyone they're ignoring.

🏦

Consolidated Market — Monopoly Pricing

Years of M&A have concentrated the structured credit data market into a single dominant franchise. Annual contracts, no self-serve, no sandbox. Budget cycles start at six figures and only go up. The underlying data? Substantially sourced from free public government filings.

Free data. 5-figure invoice. $50K – $200K / yr
📠

Proprietary Formats, Zero Developer Ergonomics

Flat file FTP drops. Proprietary field names and schemas requiring manual archaeology every vintage. No REST API, no cloud integration, no metadata discovery. Every team rebuilds the same ingestion pipeline from scratch.

Weeks of ETL per analyst
🔒

Designed for Servicers, Not Analysts

Legacy platforms were built for loan servicers and originators — not independent credit analysts, mortgage REITs, or boutique hedge funds. You pay for features you don't need and lack the query flexibility you do.

Poor fit for buy-side
Status Quo
Agency Data Platforms $50K–$200K/yr · Enterprise contract · Proprietary access only
Loan Performance Databases $30K–$150K/yr · Annual lock-in · Proprietary schema, no API
Terminal-Based Analytics $40K+/yr · Proprietary terminal dependency · Per-seat pricing

Rigorous Data. AI-Ready Schema.
A Fraction of the Cost.

TrueABS is built cloud-first on public agency sources, processed through a rigorous pipeline into a clean, MISMO-aligned semantic schema. Every field is consistent across vintages — so agents, models, and analysts can query without manual reconciliation.

📊

Rigorous, Semantic Data Model

Every field is MISMO 3.x aligned and consistent across all vintages — 10+ years of deep history. Our Format Registry tracks agency schema changes as versioned entries, not code patches. Your models query the same field names whether the vintage is 2003 or 2024. Bi-temporal tracking separates when the data changed from when our understanding of the format changed — every ingestion run is replayable and audit-proof, not just current-state.

MISMO 3.x · Bi-Temporal & Replayable · 40+ fields/loan
🤖

AI-Native by Design

Ask your data anything — in plain English. TrueABS's semantic layer means agents and analysts query the same schema without hallucinating field names or reconciling vintage formats. "90-day delinquency by servicer, GNMA, Florida, 2021 vintages?" — answered in seconds, no SQL required. Surveillance, deal summaries, cohort analysis: all AI-queryable out of the box.

Ask anything in plain English · Agentic · Semantic layer
☁️

Cloud-Native Access

Cloud-native delivery via modern open formats, data marketplace listings, and a request-driven REST API. No FTP, no proprietary client, no legacy tooling. Connect any modern analytics stack — including your AI agents and agentic workflows.

Cloud-Native · Marketplace · Request-Driven API
🔗

Full Structural Hierarchy

The complete deal→tranche→pool→loan linkage chain — traversable in a single query. REMIC deal CUSIPs down to individual loan performance, with issuer/servicer/insurer (Party) and underlying collateral (Property) linked at every loan. This is the structural depth legacy platforms charge institutional rates for. TrueABS ships it on every plan.

Deal → Tranche → Pool → Loan · Party & Property linked · REMIC aware
🏦

Practitioner-Designed, Inside Out

35+ years of front-line experience across top-tier investment banks and leading hedge funds went into every schema decision, every field name, every query pattern. We know the workflows, the edge cases, and the questions you're trying to answer — because we spent decades asking them ourselves. On the other side of that invoice.

35+ yrs institutional experience · Built for practitioners
🔄

Continuously Updated, Never Stale

Agency releases processed within 48 hours of each monthly publication. Vintage cohort consistency guaranteed through our Format Registry — field definitions stay stable even as agency schemas evolve over decades.

Monthly cadence · <48hr lag · No stale snapshots

Coverage That Matters

Built from the same public agency disclosures that underpin every competing product — processed into a single, consistent analytical schema.

FNMA

Fannie Mae Single-Family

28M+ loans
1999 – present · Monthly performance updates · Full origination cohorts

Origination characteristics (FICO, LTV, property type, occupancy, purpose, DTI) plus 12 monthly performance fields including delinquency, UPB, modification, and liquidation events.

FICO / LTV Delinquency Prepayment Modification Loss Severity UPB Property Type Vintage Cohort
FHLMC

Freddie Mac Single-Family

55M+ loans
1999 – present · Monthly performance · Largest agency dataset

The deepest single-agency loan-level dataset available. 27 monthly performance fields including zero-balance codes, repurchase flags, and MI recoveries — more granular more granular than what legacy platforms expose to non-enterprise subscribers.

Credit Score Zero-Balance Codes MI Recovery Repurchase Flags ARM Rate Resets Servicer Name MSA
GNMA

Ginnie Mae (FHA / VA / USDA)

11M+ loans
Monthly loan-level performance · Pool factors · REMIC / MBS disclosure

Loan-level performance and payment history for FHA, VA, and USDA programs, plus pool-level factor updates, REMIC tranche allocations, and servicer/issuer attributes. Government mortgage analytics without a proprietary terminal or platform dependency.

Delinquency Pool Factors FHA / VA / USDA REMIC Disclosure Issuer / Servicer WAC / WAM
Non-Agency

Non-Agency RMBS (SEC EDGAR)

Coming Soon
2015+
SEC ABS-EE machine-readable disclosure · Deal tranche mapping · Waterfall terms

SEC-mandated ABS-EE XML loan-level data for non-agency RMBS issuances post-2015, linked to deal and tranche CUSIPs extracted from 424B5 prospectuses via an LLM extraction pipeline. Deal structure, waterfall triggers, and collateral stratifications — in active development.

ABS-EE XML 424B5 Extraction CUSIP Linkage Waterfall Terms Tranche Structure
Enrichment

Reference Data & Climate Risk Overlays

Coming Soon
6 datasets
FHFA HPI · FIO Insurance · FRED · ACS Demographics · FEMA NRI · HMDA

Pair FIO's ZIP-level homeowners insurance nonrenewal data with agency loan performance — joined at state level directly, MSA level at the pool, and statistically inferred at the loan level. Estimate current LTV via FHFA HPI, overlay FEMA hazard scores, benchmark against HMDA originations. No other platform at this price combines these layers.

FHFA HPI (current LTV) FIO Insurance Risk FEMA Hazard Index FRED Rate Series ACS Demographics HMDA Originations
Graph Engine

Recursive Tranche & Pool Resolution

Coming Soon
Ontology · Any Depth
Tranche · Platinum Pool · Conventional Pool · arbitrary nesting

Resolve any tranche — nested across sub-tranches and platinum pools to any depth — into a flat leaf-pool weight table that always sums to 1.0, then join straight into loan-level FICO, CPR, and geographic concentration. One call, no incumbent exposes this depth of structure today.

resolve_tranche() Multiplicative Weight Accumulation MCP / Agentic Tool
Full Structural Depth REMIC Deal Tranche / Class Pool(s) Loan-Level
Linked At Every Loan Party (Issuer / Servicer / Insurer) Property (Collateral)

Complete deal-to-loan traversal, queryable in a single call — with issuer/servicer and underlying collateral attached at the loan level. This is the structural depth legacy platforms charge institutional rates for — TrueABS ships it on every plan.

Even the modern cloud-native alternatives stop at the pool level — no REMIC structure, no tranche→pool→loan linkage, no non-agency coverage. That ceiling is exactly where TrueABS starts.

Query 100M+ Loans. Ask in Plain English.

SQL, Python, or natural language — TrueABS is built so analysts and AI agents can query without schema archaeology. A request-driven, metadata-discovery architecture means no batch jobs, no waiting, no data wrangling.

trueabs — query console

TrueABS Demo — Coming Soon

Full platform walkthrough — SQL, natural language queries, and API access in under 5 minutes.

SELECT orig_vintage_year, AVG(fico_score) AS avg_fico, AVG(orig_ltv_ratio) AS avg_ltv, SUM(CASE WHEN dq_status >= 60 THEN upb_at_issuance END) AS dq60_upb
FROM loan_performance_all
WHERE agency = 'FNMA' AND orig_purpose = 'Purchase' AND orig_vintage_year BETWEEN 2020 AND 2023 GROUP BY 1 ORDER BY 1;

-- ✓ 3 rows · 0.18s · 27.2M loans scanned
"30 seconds. Not 3 weeks of data wrangling." What analysts say when they first query 100M+ loans on TrueABS.

Two Ways to Run TrueABS Today

Every path runs on the same Format Registry — 10+ years of agency format history no competitor has reconstructed, bi-temporal and audit-proof by design. Pick the deployment model that fits your team, then the coverage tier that fits your book.

Deployment Model Starter Core Full Agency + Enrichment Enterprise
Managed Cloud Delivery Same operating model as Managed Snowflake Delivery — we run the Dagster pipelines, nothing runs on your infrastructure. We land Parquet in your own S3 / ADLS / GCS, then materialize staging and semantic tables on your chosen warehouse (Redshift, Synapse, BigQuery, Databricks). Unlike Snowflake's single COPY INTO target, that's a bespoke load process per platform — the reason this tier costs more. $15K+/yrGNMA only $35K+/yrGNMA + FNMA $50K+/yrAll 3 agencies + FHFA HPI / FIO enrichment $75K+/yrCustom agency coverage + semantic layer
Future Hosted TrueABS Query via REST API — we'd operate the pipeline, the data, the uptime SLA, and the hosting compute ourselves, same cost burden as Managed Cloud Delivery rather than Managed Snowflake Delivery (where the customer's own Snowflake compute covers it). Priced the same as Managed Cloud Delivery as a result. Floors, not final. Planned, not yet available. $15K+/yrGNMA only $35K+/yrGNMA + FNMA $50K+/yrAll 3 agencies + FHFA HPI / FIO enrichment $75K+/yrCustom agency coverage + semantic layer

Managed Snowflake Delivery and Managed Cloud Delivery are available now. Hosted TrueABS is a planned future offering — a hosted query layer with an uptime SLA is the heaviest lift of the three, and it's not where we're starting.  Talk to us about which fits →

Get Access to TrueABS Before We Launch

We're onboarding a limited cohort of private beta users. Rigorous data, AI-native queries, and institutional coverage — waitlist subscribers receive 3 months free on any paid tier.

Early access includes 3 months free on any paid tier.

You're on the list!

We'll be in touch within 2 business days.