datacendia-core

# Datacendia — Cryptographic Proof for Every AI Decision Your Enterprise Makes
# https://app.datacendia.com
# Contact: sales@datacendia.com

## What is Datacendia?

Datacendia is open-source decision governance infrastructure (DDGI). You pose a question — about a hire, a deal, a policy change, a risk — and a panel of domain-expert AI agents (CFO, CISO, Legal Counsel, Ethics Officer, Red Team) forces adversarial deliberation from every angle. Every argument, every dissent, every vote is permanently recorded and cryptographically signed. The output is a forensic-grade, independently verifiable decision record, not a chat log. Built on the DDGI Open Standard.

## How is Datacendia different from ChatGPT?

ChatGPT is a conversation — one model, one perspective, no record. Datacendia is a governed decision:
- Multiple domain-expert AI agents argue from different perspectives
- Disagreements are recorded as first-class artifacts (structured dissent)
- Every response is cryptographically signed (Ed25519), hashed into a Merkle tree, and RFC 3161-timestamped
- The output is a forensic-grade, independently verifiable evidence bundle with full chain of custody
- A human must sign responsibility for every AI-assisted decision

## Two Editions (Commercial Open-Source / COSS)

- **Community Edition**: Open Source (Apache 2.0). Full platform — frontend, backend, deliberation engine, cryptographic infrastructure. Self-hosted, single node. All inference providers (Ollama, OpenAI, Together AI, Anthropic Claude, Google Gemini). Community support via GitHub.
- **Enterprise Edition (Sovereign)**: Air-gapped/sovereign deployment. Post-quantum KMS integration. CendiaInsure liability coverage. Custom integrations + dedicated success team. SSO/SAML, RBAC, multi-tenant. Annual license + support. Contact sales.

## Key Capabilities

- Multi-agent AI deliberation with domain-expert agents (CFO, CISO, Legal, Ethics, Red Team)
- Ed25519 cryptographic signatures on every AI response
- Merkle tree hashing for tamper-proof decision chains
- RFC 3161 timestamping for legal admissibility
- Regulator's Receipt™ — one-click forensic-grade, independently verifiable evidence export
- Structured dissent recording — minority opinions are first-class artifacts
- CendiaDCII™ — Decision Crisis Immunization Infrastructure with 9 governance primitives
- CendiaIISS™ — Institutional Immune System Score for organizational resilience
- CendiaMediaAuth™ — Synthetic media authentication with C2PA support
- CendiaCascade™ — Second and third-order consequence analysis (butterfly effect detection)
- CendiaNotary™ — Cryptographic signing authority with AWS KMS, Azure Key Vault, HashiCorp Vault
- CendiaVault™ — Unified evidence storage with legal hold and retention policies

## Compliance Frameworks

Datacendia maps to:
- EU AI Act (Article 14 human oversight)
- NIST AI RMF (AI Risk Management Framework)
- ISO 42001 (AI Management System — aligned, awaiting certification)
- SOC 2 Type II (aligned, awaiting certification)
- ISO 27001 (aligned, awaiting certification)
- GDPR Article 22 (automated decision-making)
- HIPAA
- SR 11-7 (model risk management)
- ABA Formal Opinion 512 (lawyers using AI)

## Technical Stack

- Frontend: React + TypeScript + TailwindCSS
- Backend: Node.js + Express + Prisma
- Inference: Ollama, OpenAI, Together AI, Anthropic Claude, Google Gemini (BYO-Key) / NVIDIA Triton + NIM (Sovereign)
- Cryptography: Ed25519 signatures, Merkle trees, RFC 3161 timestamps
- License: Apache 2.0 open source
- Languages: 26 supported (English, Spanish, French, German, Japanese, Chinese, Arabic, Hindi, and 18 more)

## Company

- **Name**: Datacendia, LLC
- **Founded**: 2024
- **Program**: NVIDIA Inception Program member
- **License**: Apache 2.0
- **Website**: https://datacendia.com
- **Platform**: https://app.datacendia.com
- **GitHub**: https://github.com/datacendia

## FAQ

Q: Is my data safe?
A: Datacendia Core is cloud-hosted with all data encrypted at rest and in transit. The Sovereign Edition deploys on-premises in air-gapped environments — your infrastructure, your keys, your control.

Q: What LLMs does it support?
A: Core uses cloud LLMs via OpenAI-compatible APIs (GPT-4o, Groq, Together AI). Sovereign supports local models via Ollama (Llama 3, Mistral, Phi-3, Gemma) and NVIDIA Triton/NIM.

Q: Is it open source?
A: Yes. Apache 2.0 licensed. The full platform is open source. Enterprise features are about support and SLAs, not locked functionality.

Q: Can I verify the cryptographic signatures?
A: Yes. Every decision packet includes Ed25519 signatures and RFC 3161 timestamps verifiable with standard openssl commands. No proprietary tools or vendor dependency.

Q: What is the Regulator's Receipt?
A: The Regulator's Receipt™ packages the complete decision record — all agent arguments, votes, dissent, Merkle proofs, signatures, timestamps, and chain of custody — into a single forensic-grade, independently verifiable evidence bundle.

Q: How long does setup take?
A: Core is cloud-hosted — sign up and run in seconds. Sovereign takes five minutes with Docker, or 1-2 days for enterprise deployments with SSO and air-gap configuration.