OpenViking

OpenViking is a self-evolving context database for AI agents. It unifies agent memory, knowledge RAG and skills into one context layer, so agents remember what they did, retrieve what they know and reuse skills across sessions.

Key Features

• Unified context: memory, knowledge and skills in one system.
• Self-evolving: context improves as agents work.
• RAG built in: retrieve relevant knowledge at inference time.
• Agent-native: designed for LLM agents, not just chat.
• Open source under AGPL-3.0.

Getting Started

Run OpenViking locally or in your stack, point your agent framework at it, and start persisting agent context. See openviking.ai for docs.

Pricing

Free and open source (AGPL-3.0).