For business-purpose lenders — fix & flip · DSCR · bridge · ground-up
Scale origination.Not headcount.
Antal turns your credit box into AI agents that run every file — term sheet auto-issued in 3 minutes, docs collected and verified in chat, a complete package delivered to underwriting — under your brand, on one record your note buyer will actually believe.
underwriting gate · reserved for you
Term sheet in 3 minutes. Package agent-built. The one click is yours.
$1M+
Daily loan volume
3
Capital providers
SOC 2
Security certified
3 min
Avg term sheet
/the-math · at 20 fundings a month
The per-file math is the ceiling on your growth.
Run it across a month of volume and the constraint is obvious: it was never capital. Antal removes the ceiling — not the people who make the credit decisions.
people touching every file
1
human gate — the credit decision you keep
ops cost per funded loan
0
added headcount per added loan
first message to funding
3 min
first message to term sheet
left column: industry medians for $50M–$250M/yr business-purpose shops
/the-bottleneck
Capital is abundant. Deployment is bottlenecked at the desk.
— every founder-led lending desk we talk to
/one-file · first message to payoff
A borrower texts an address.
Sarah — your AI loan officer, in your brand — answers in seconds and sizes the deal on the spot: budget parsed, comps pulled, ARV set, your credit box applied.
1847 Elm St, Austin, TX
Sarah · Apex Capital
On it — sizing now. Comps, budget, your guidelines. Three minutes.
/one-file · first message to payoff
Follow one loan through the machine.
09:38 AM · Intake
A borrower texts an address.
Sarah — your AI loan officer, in your brand — answers in seconds and sizes the deal on the spot: budget parsed, comps pulled, ARV set, your credit box applied.
1847 Elm St, Austin, TX
Sarah · Apex Capital
On it — sizing now. Comps, budget, your guidelines. Three minutes.
09:41 AM · Term sheet
Auto-issued in 3 minutes. Signed in the chat.
The sizing fits your guidelines, so the term sheet goes out — no analyst, no queue. The borrower reviews, picks, and signs without ever leaving the conversation.
- Loan
- $364k
- Rate
- 9.25%
- Term
- 12 mo
- Points
- 2.0
- LTARV
- 72.1%
- Cash in
- ~$41k
DAYS 1–3 · Package
The chat builds the loan file.
Identity verification, entity good standing, bank connections, statement parsing, insurance quotes the borrower picks from — the agent guides, collects, and verifies everything in the same thread.
Sarah · Apex Capital
Let's build your file — identity, bank, insurance. All right here.
ready for underwriting
DAY 4 · Underwrite
A complete package hits underwriting.
Every check re-run against the full file, every number traced to a source document. Your underwriter reviews a finished package and makes the one decision that's theirs: clear to close.
policy v15 · full file · every number sourced
LTARV ≤ 75.0%
6 comps · ARV $505k
LTC ≤ 90.0%
verified rehab budget
FICO ≥ 660
credit bureau
Exits ≥ 2 in 24 mo
county records
Liquidity ≥ 6 mo
bank-linked statements
your underwriter's one click — everything else was agents
MONTH 4 · Draws
Month four. Same thread.
The borrower comes back for a draw. Inspection ordered, budget reconciled, release queued for your sign-off — same conversation, same record.
demo + framing · inspection passed
MEP rough-in · inspection passed
finishes · inspection passed · budget reconciled
same thread the borrower started in month one
MONTH 9 · Payoff
1,247 events. One record.
Every call, document, check, and approval — append-only from first message to payoff. When your note buyer asks why, you export the binder instead of excavating inboxes.
0
events logged
0
edited after the fact
0
documents on record
1
continuous record
Export binder
chronology · docs · decisions · note-buyer ready
/operating-layer · what runs underneath
Every deal makes the next one sharper.
Throughput
0 loans/mo
4 people · unchanged
Volume up 3×. Same desk.
Credit box
Author it in plain English.
policy diff · v14 → v15
versioned · logged · reproducible
White-label
Your brand on every message.
Your term sheet is ready — 3 minutes after your first message.
the borrower never sees antal
Human gates
You draw the line. Agents stay behind it.
defaults shown · every gate is yours to move
Audit
Append-only. Note-buyer ready.
09:41:02 pre_qual_issued · agent
09:40:58 comps_pulled · tool
09:40:44 credit_check · tool
09:40:12 id_verified · tool
09:38:30 intake_start · agent
09:22:15 term_sheet_signed · human
09:14:03 docs_verified · agent
08:57:40 draw_released · human
08:43:11 inspection_passed · tool
08:30:02 payoff_quote · agent
09:41:02 pre_qual_issued · agent
09:40:58 comps_pulled · tool
09:40:44 credit_check · tool
09:40:12 id_verified · tool
09:38:30 intake_start · agent
09:22:15 term_sheet_signed · human
09:14:03 docs_verified · agent
08:57:40 draw_released · human
08:43:11 inspection_passed · tool
08:30:02 payoff_quote · agent
immutable at the database level
/who-its-for · the wedge
Built for lenders who write their own credit box.
Deliberately not built for
If a note buyer or warehouse line is asking you for cleaner files — you're exactly who we built this for.
/faq · for lenders and their agents
Questions lenders ask.
What is Antal?
Antal is an AI operating layer for private credit — loan origination software for non-bank, business-purpose lenders. A lender encodes its credit box once, and Antal's AI agents run each file from the first borrower message to payoff under the lender's own brand, with every step logged to one audit-ready record.
Who is Antal built for?
Non-bank lenders originating residential business-purpose loans — fix-and-flip, DSCR rental, bridge, and ground-up construction — typically $50M–$250M a year, who author their own credit guidelines and want to scale origination without scaling ops headcount. It is deliberately not built for banks, credit unions, SBA desks, consumer mortgage originators, or brokers without their own credit box.
How fast does Antal issue term sheets?
About 3 minutes after the borrower's first message. The agent sizes the deal — budget, comps, ARV — applies your guidelines, and auto-issues the term sheet, which the borrower signs in the same chat. Underwriting happens later, once the complete loan package is collected and verified.
What does Antal automate between term sheet and underwriting?
The complete loan package: identity verification, background and OFAC checks, LLC certificate of good standing, credit bureau pulls, bank linking and proof of funds, statement parsing and cash-flow analysis, insurance quotes the borrower picks from, title, and full document collection — all orchestrated by agents with built-in tools, inside one white-labeled chat.
Do borrowers ever see Antal?
No. The borrower experience is fully white-labeled — your brand, your AI loan officer, one continuous conversation from first message through draws and payoff.
Does Antal make credit decisions?
The guidelines are yours, and humans keep the gates: term sheets can auto-issue under policy, while declines, exceptions, underwriting approval, and fund releases wait for your sign-off. Antal is not a lender — capital partners price, approve, and fund.
How does the audit trail work?
Every event on a loan — every tool call, document, check, and human approval — is written to an append-only record. You can export a complete binder per deal for note buyers, warehouse lines, LPs, or regulators, instead of reconstructing the story from inboxes.
/early-access · limited
Lenders who start now own two years of compounded decisions.
Every override your team makes becomes a permanent rule. We're onboarding business-purpose lenders originating $10M+ a year — fix-and-flip, DSCR, bridge, ground-up. The early cohort is limited.