Cap Orbit vs Apers: Full Deal Lifecycle vs Underwriting Autopilot

Credit where due

A capable engine, and one specific edge.

This is the rare comparison in the series where both products were built for the same team. Apers describes itself as an AI autopilot for institutional CRE investors: the XL-2 engine takes the deal documents, the OM, the rent roll, the T-12, the appraisal, and generates an Excel workbook with citations down to the cell, and a dedicated extraction engine reconciles the figures that disagree across sources.

The coverage list runs from waterfall and debt sizing through tax-credit structures, including 4% and 9% LIHTC executions, with asset management, fund reporting, and lease analysis alongside, across multifamily, office, retail, logistics, and self-storage. Tax-credit structures are Apers’s clearest edge, and a team whose book runs on LIHTC should weight that.

Apers publishes its own comparison pages; read them alongside this one. The rest of this page is about what happens to the deal after the workbook ties out, which is where the two products stop resembling each other.

Where Cap Orbit wins

Most of a deal happens after the workbook ties out.

Start with how the work gets done. The Cap Orbit terminal has the run of the deal file: one instruction reads the offering memo, the rent roll buried in a workbook tab, the T-12, and the loan documents at once, normalizes the statement, builds the model, and stages the memo, with the analyst approving each consequential step. It is the difference between asking a question about an attached document and getting back the workbook, memo, and record: a real workbook with live formulas, the memo in the house format, the deck, the bound and bookmarked PDF, all written back into the deal file.

The model is the start of the argument, not the end of it. When the deal goes to committee, Cap Orbit drafts the memo in the format and voice the committee already reads: point it at the firm’s filed memos and it studies them and writes in that voice, with the section-by-section outline approved before a word is drafted. Every figure in the prose comes from the model’s computed cells or a cited document, the memo reads the model and never writes it, and a number the documents do not contain stays a flagged blank for the analyst to fill. The Apers coverage lists IC memo generation; nothing in their public materials describes outline approval or house-voice study, so put the how to them directly.

Then the deal signs, and the numbers start moving. Cap Orbit abstracts the executed agreement, every amendment, and the conditions to funding into one source-cited reference, reconciles the settlement statement against the contract, the loan, and the underwrite, flags each variance with its cause, and writes the trued-up going-in basis back into the model, so the deal of record is the deal that actually closed. As of mid-2026, nothing in the public Apers materials describes this stretch between signing and funding.

The hold is where the shape of the record starts to pay. Apers lists asset management and fund reporting in its coverage. Cap Orbit is specific about how that record is kept: underwrite, budget, forecast, and actuals on one append-only lineage, each period a new dated entry, restatements filed beside the original and never over it, and covenant standing read from the loan documents with each test cited to its section, an internal read for the team, not a lender certificate. Five years in, the question is not whether a number exists but whether you can prove what it was and when it changed.

The deployment question

Where the deal documents actually sit.

For an institutional buyer the second conversation is always with the security team, so here is the posture plainly. In Cap Orbit, each firm runs walled off with its own database and its own document storage; broad default access is revoked, and writer privileges reach that firm’s database alone. Within the firm, each deal runs sealed in its own dedicated compute with only its own files attached. No standing credential sits in the data path: each request is re-verified against the firm’s identity provider and issued a pass good for fifteen minutes. Single sign-on is the only door, and customer files, prompts, outputs, and templates are never used to train any model.

On deployment, two tiers. Pro is the managed tier for funds and deal teams of up to 50 people: every firm isolated on its own dedicated resources, up and running with live deals within 24 hours. Enterprise deploys the same platform into the firm’s own AWS account, with single sign-on, private connectivity, customer-managed encryption keys, and bring-your-own-key for the inference models, built against your security and architecture review.

The public Apers materials describe what the engine produces, not where the work sits. As of mid-2026 they do not advertise per-firm isolation, a path into the customer’s own cloud account, or a training-data commitment, which is not the same as lacking them. Put the same three questions to both vendors: who can see the deal file, who can write into the environment, and what, if anything, trains on your documents.

The buyer’s read

Buy for the phase, or buy for the hold.

If the job you are hiring for ends at the underwrite, Apers is worth testing, and a book that runs on 4% and 9% LIHTC executions is exactly where it specializes.

Cap Orbit is the pick when the firm holds what it buys. The workbook still has to tie out, but so does the memo against the model, the settlement statement against the contract, and the fifth-year actuals against the underwrite the committee approved. One deal file carries the sources, the model, the drafts, and the version history; the deal briefing carries the case into every future conversation; and no assumption writes into the model until the analyst accepts it, because the team runs the numbers and the recommended action stays with your team.

The cleanest way to decide is also the cheapest: pick one live deal and run it through both. On the Pro tier the firm is up and running with live deals within 24 hours, so the test costs a day to start, not a procurement cycle. Where the products differ, you will see it on your own documents, not in either vendor’s comparison pages.