AIVXY Studio

AIVXY is a privacy-first AI platform providing local-first inference, encrypted memory, and agent orchestration for secure, on-premises workflows.It supports local model runtimes (Ollama) and optional cloud connectors (OpenAI, Anthropic, Gemini) with 90+ API endpoints and a hand-authored OpenAPI 3.1 spec for developer integration.Built in Rust and open source (MIT), aivyx includes 46+ built-in tools, 9 specialist agents, DAG orchestration, reflection loops, and human-in-the-loop checkpoints for reliable agent workflows.

Data protection features include per-tenant HKDF key derivation, ChaCha20-Poly1305 authenticated encryption, HMAC audit logging, RBAC, OIDC SSO, and prompt-injection mitigations for enterprise compliance.

Multimodal capabilities cover real-time STT/streaming TTS voice agents, image and document understanding (PDF, XLSX, CSV), multimodal memory with embeddings, and RAG-style retrieval from an encrypted knowledge graph.

Operational tooling includes CLI, TUI, desktop installers, webhooks/workflows with HMAC signing, Kubernetes Helm charts for scaling, PostgreSQL/Redis backends, and usage/cost governance controls.Target users include developers, security and compliance teams, and enterprises needing locally runnable, auditable AI agents with strong encryption and integration options.

Key Features

• Local-first inference with support for local model runtimes (Ollama) and optional cloud connectors (OpenAI, Anthropic, Gemini) via 90+ API endpoints and an OpenAPI 3.1 spec
• Encrypted memory and data protection with per-tenant HKDF key derivation, ChaCha20-Poly1305 authenticated encryption, and HMAC audit logging
• Agent orchestration with built-in tools and specialist agents, DAG orchestration, reflection loops, and human-in-the-loop checkpoints
• Multimodal capabilities including real-time STT and streaming TTS voice agents, image and document understanding (PDF, XLSX, CSV), multimodal embeddings, and RAG retrieval from an encrypted knowledge graph
• Operational and deployment tooling: CLI/TUI/desktop installers, webhooks/workflows with HMAC signing, Kubernetes Helm charts, and PostgreSQL/Redis backends

Use Cases

• Create on‑premises, privacy-first customer support agents with aivyx that run local model inference and encrypted agent memory (ChaCha20-Poly1305, HKDF), orchestrate multi-step responses via DAG workflows, accept multimodal inputs (voice, images, documents) and expose secure APIs/OpenAPI for enterprise integration with RBAC and SSO
• Build an auditable, open-source knowledge assistant for sensitive internal data using aivyx to perform encrypted knowledge retrieval and multimodal document understanding, keep all processing local for data residency, log and audit agent actions, and enforce access controls with RBAC and SSO
• Develop automated compliance and document-processing pipelines using aivyx agent orchestration to extract data from contracts, invoices and voice notes, run workflows locally with encrypted memory, integrate via extensive APIs/OpenAPI, and maintain enterprise-grade security and auditability for regulated environments

Who is it for?

• Developers
• Security teams
• Compliance teams
• Enterprise it teams
• Devops/sre engineers