# Divinci AI

> Release management and quality assurance for custom language models.
> Divinci gives enterprises a structured path from a fine-tuned or
> retrieval-augmented model to a production release: automated evaluation,
> retrieval quality control, compliance evidence, and instant rollback.

Divinci is a platform, not a model vendor. Customers bring their own models and
content; Divinci handles ingestion, retrieval, evaluation, release gating, and
the audit record that regulated industries need. Assistants built on it are
embedded on customer sites, reached by phone, or driven by other agents over
MCP.

## For agents

- [MCP server description](https://mcp.divinci.app/info): tools, transports and OAuth endpoints
- [Agent skills index](https://divinci.ai/.well-known/agent-skills/index.json): how to connect, query RAG, and embed an assistant
- [auth.md](https://divinci.ai/auth.md): how an agent registers and authenticates
- [API catalog](https://divinci.ai/.well-known/api-catalog): RFC 9727 service directory
- [A2A agent card](https://divinci.ai/.well-known/agent-card.json): agent-to-agent discovery
- [SDK and API documentation](https://sdk.divinci.ai/)
- [Platform status](https://divinci.ai/api/status)

## Product

- [Release management](https://divinci.ai/release-management/): staged rollout, approval gates, rollback
- [Quality assurance](https://divinci.ai/quality-assurance/): automated evaluation before a model reaches users
- [WWW-RAG](https://divinci.ai/www-rag/): open-web crawling and retrieval
- [AutoRAG](https://divinci.ai/autorag/): managed retrieval pipelines
- [RAG routing](https://divinci.ai/rag-routing/): choosing the right knowledge source per query
- [RAG Arena](https://divinci.ai/rag-arena/): head-to-head retrieval evaluation
- [Voice agents](https://divinci.ai/voice-agents/): phone and realtime speech assistants
- [Hermes agents](https://divinci.ai/hermes-agents/): long-running autonomous agents
- [Local inference and privacy](https://divinci.ai/local-inference-privacy/): in-browser and on-device models
- [Pricing](https://divinci.ai/pricing/)

## Trust and compliance

- [Security](https://divinci.ai/security/)
- [Compliance](https://divinci.ai/compliance/): EU AI Act, GDPR, HIPAA, NIST AI RMF
- [AI safety](https://divinci.ai/ai-safety/)
- [Privacy policy](https://divinci.ai/privacy-policy/)
- [Data processing agreement](https://divinci.ai/data-processing-agreement/)
- [Terms of service](https://divinci.ai/terms-of-service/)
- [Accessibility](https://divinci.ai/accessibility/)
- [security.txt](https://divinci.ai/.well-known/security.txt)

## Developers

- [API overview](https://divinci.ai/api/)
- [CLI](https://divinci.ai/cli/)
- [Documentation](https://divinci.ai/docs/)
- [Tutorials](https://divinci.ai/tutorials/)
- [Changelog](https://divinci.ai/changelog/)
- [Roadmap](https://divinci.ai/roadmap/)

## Company

- [About](https://divinci.ai/about/)
- [Careers](https://divinci.ai/careers/)
- [Press](https://divinci.ai/press/)
- [Contact](https://divinci.ai/contact/)
- [Support](https://divinci.ai/support/)
- [Blog](https://divinci.ai/blog/)

## Optional

- [Full site content](https://divinci.ai/llms-full.txt): every indexed page with its summary
- [Sitemap](https://divinci.ai/sitemap.xml): all 700+ URLs including translations
- [Brand assets](https://divinci.ai/brand/)