# Anteroom

> Anteroom is a structured legal-review surface for AI deployment work. Four primary-source-cited modules: AI launch readiness (regulatory corpus across 15 jurisdictions), AI partnership architecture (tracked deals and structural patterns), AI vendor commitments (commercial-term tracking across 15 vendors), and frontier-lab safety frameworks (RSP, Preparedness Framework, FSF version diffs).

Not legal advice. Not a substitute for counsel review. Anchored to public 8-K filings, regulator press releases, framework PDFs, and court dockets. Built by Marcus Harjani.

## Core surfaces

- [Homepage](https://anteroom.so/): Project overview, three reads on AI deployment, two interactive tools.
- [About](https://anteroom.so/about): Who built Anteroom and the premise behind it.
- [Methodology](https://anteroom.so/methodology): How Anteroom curates, classifies, and verifies its corpus.

## AI launch readiness

- [Corpus](https://anteroom.so/corpus): Every regulatory provision Anteroom tracks. Filterable by lens and jurisdiction. Reviewer-attributed.
- [Corpus changelog](https://anteroom.so/corpus/changelog): Provision-level diffs by date.
- [Blog: Why a form beats a chat](https://anteroom.so/blog/why-a-form-beats-a-chat): The argument for structured AI legal review.

## AI partnership architecture

- [Partnerships index](https://anteroom.so/partnerships): Every tracked AI strategic partnership. Searchable, filterable by structure type and status.
- [Patterns index](https://anteroom.so/patterns): The structural shapes AI partnerships fall into, with kill-list moves and precedent anchors.
- [New partnership profile](https://anteroom.so/partnerships/new): The interactive tool. Describe a deal, get matched patterns, kill-list moves to watch, similar tracked deals, precedent to consider.

## AI vendor commitments

- [Commitments matrix](https://anteroom.so/commitments): 15 AI vendors tracked across 7 commercial commitments (AUP, no-train default, output ownership, IP indemnity program, retention default, deprecation notice, safety framework version). Primary-source-cited.
- [Commitments changelog](https://anteroom.so/commitments/changelog): Detected changes in vendor commitments by date.

## Frontier-lab safety frameworks

- [System Cards index](https://anteroom.so/system-cards): Anthropic, OpenAI, and Google DeepMind tracked. Framework version timelines (RSP, Preparedness Framework, FSF) with key changes per version. Per-model system cards with eval categories.

## Signals

- [Signals feed](https://anteroom.so/signals): What Anteroom heard from regulators and practitioners. Classified by a small model, scored by materiality, mapped to the corpus when applicable.

## Optional

- [Blog](https://anteroom.so/blog): Weekly posts. Pillar-cluster discipline across the four modules. Anchored to primary sources.
- [RSS](https://anteroom.so/blog/rss.xml): Blog feed for RSS readers.
- [Privacy](https://anteroom.so/privacy): Plain-language overview of what Anteroom stores.
- [Terms](https://anteroom.so/terms): Plain-language terms of use.