# Decrypted Matrix Nexus > An uncensored AI research assistant connected to the Decrypted Matrix vault of primary-source documents: court records, filings, FOIA releases, and declassified material. It quotes what it retrieves and marks what it cannot verify. ## What this is A browser chat app at www.decryptedmatrix.ai with two assistants, switched from the composer: - NEX — research and legal work, grounded in the vault. Has retrieval, legal search, and scholar tools, and cites what it uses. - NEXUS — health, sovereignty, and personal work. No vault, no tools. There is no hardware, GPU time, hourly compute, prepaid credits, or self-hosting. It is a browser app and a subscription. ## Plans - Free: 8 messages per week, no credit card. Counted over a rolling 7 days, not a calendar week. - NEX Pro: $20/month. Unlimited messages on both assistants, unlimited document attachments, voice chat, earliest access to new vault material. - Pro fair-use ceiling: 500 messages per calendar month. It exists to bound a scripted or compromised account, not to ration ordinary use. - Document attachments in chat: 1 per 7 days on the free tier, unlimited on Pro. - Billing is monthly through Stripe. Cancel any time from /account; Pro runs to the end of the period already paid for. No prorated refunds for a partial month. ## Documents Chat accepts PDF, DOCX, and JPG or PNG photos up to 10 MB. Scans and photos are read with OCR server-side, so a phone picture of a paper document works. The separate document review tool accepts PDF and DOCX only. ## Citations, and what unverified means When a case or filing cannot be confirmed against the public record it is marked unverified, which means it could not be checked, not that it is false. Case-law lookup covers United States courts, so a foreign judgment or a document absent from public databases comes back unconfirmed even when it is genuine. ## API Two JSON endpoints over the vault, documented at /docs/api: - POST /api/v1/search — similarity search, 60 requests/minute per key. - POST /api/v1/verify — splits text into claims and returns vault sources per claim, 10 requests/minute per key. Bearer key in an Authorization header. Keys are issued by hand to Pro customers, not self-serve: email support@decryptedmatrix.ai. Each call spends one message from the same monthly allowance as chat. ## Pages - [Home](https://www.decryptedmatrix.ai/) - [Pricing](https://www.decryptedmatrix.ai/pricing) - [FAQ](https://www.decryptedmatrix.ai/faq) - [Vault API reference](https://www.decryptedmatrix.ai/docs/api) - [Document review, how it works](https://www.decryptedmatrix.ai/review/how-it-works) - [Blog](https://www.decryptedmatrix.ai/blog) - [Security](https://www.decryptedmatrix.ai/security) - [Privacy policy](https://www.decryptedmatrix.ai/privacy) - [Terms of service](https://www.decryptedmatrix.ai/terms) ## Contact support@decryptedmatrix.ai ## Articles - [Judge Rules DOD Unlawfully Retaliated Against Anthropic: A Landmark Victory for Uncensored AI](https://www.decryptedmatrix.ai/blog/dod-retaliation-anthropic-uncensored-ai-ruling) — 2026-09-02 - [Grok's Encrypted Prompt Injection: Why AI Censorship Controls Are Architecturally Doomed](https://www.decryptedmatrix.ai/blog/grok-encrypted-prompt-injection-ai-censorship-fails) — 2026-08-26 - [FOIA Document Analysis with AI: A Practical Guide to Declassified Records](https://www.decryptedmatrix.ai/blog/foia-document-analysis-ai-declassified-records) — 2026-08-19 - [Watermarking as Control: How AI Provenance Standards Are Building the Infrastructure for Content Censorship](https://www.decryptedmatrix.ai/blog/ai-watermarking-provenance-censorship-infrastructure) — 2026-08-12 - [Trump's AI Testing Exemption: How Regulatory Carve-Outs Accelerate Uncensored AI Development](https://www.decryptedmatrix.ai/blog/trump-ai-testing-exemption-open-models-uncensored) — 2026-08-05 - [AI Alignment Theater: Why Safety Claims Are Unverifiable and Deception Is Inevitable](https://www.decryptedmatrix.ai/blog/ai-alignment-theater-deceptive-models-safety-claims) — 2026-07-29 - [OpenAI's Hugging Face Breach: Why Autonomous AI Systems Demand Decentralized Infrastructure](https://www.decryptedmatrix.ai/blog/openai-hugging-face-breach-autonomous-ai-decentralization) — 2026-07-22 - [Bonsai's 1-Bit Quantization: How 27B Models Now Run on Your iPhone](https://www.decryptedmatrix.ai/blog/bonsai-1-bit-quantization-27b-iphone-local-ai) — 2026-07-17 - [Zero-Trust AI: Preventing Autonomous Systems from Going Rogue](https://www.decryptedmatrix.ai/blog/zero-trust-ai-autonomous-system-security) — 2026-04-29 - [AI Agents Unleashed: The Ethical Tightrope of Machine Autonomy](https://www.decryptedmatrix.ai/blog/ai-agents-ethical-autonomy-risks) — 2026-04-22 - [Zero-Knowledge Proofs: The Privacy Shield for Uncensored AI](https://www.decryptedmatrix.ai/blog/zero-knowledge-proofs-ai-privacy-shield) — 2026-04-15 - [Viatoris: Cracking the Code of Enterprise AI Accountability](https://www.decryptedmatrix.ai/blog/viatoris-enterprise-ai-accountability) — 2026-04-09 - [AI security is broken: authentication failures, GPU attacks, and what actually helps](https://www.decryptedmatrix.ai/blog/ai-security-vulnerabilities-uncensored-risks) — 2026-04-03 - [Welcome to Decrypted Matrix AI](https://www.decryptedmatrix.ai/blog/welcome-to-decrypted-matrix) — 2026-03-15