Features
- Persistent user memory shared across LLMs such as GPT-4, Claude and Gemini
- Hybrid search combining vector embeddings and knowledge graphs
- Forget layer to remove specific memories or topics from the index instantly
- Automated knowledge synthesis that summarizes documents and flags contradictions
- Cross-agent sync so one agent's knowledge is available to others
- Sub-50ms retrieval for complex memory queries
- Simple API that integrates with stacks like LangChain and CrewAI
Pricing Details
💵 Pricing & Plans (March 2026 Status): Memdex typically scales based on the amount of "Knowledge Points" (indexed chunks of data) and retrieval frequency. Plan Price (Approx.) Key Features Hobbyist Free Up to 10k Knowledge Points; 100 daily queries; Community support. Developer ~$29 / mo 1M Knowledge Points; Unlimited queries; Knowledge Graph features. Team ~$149 / mo 10M Knowledge Points; Shared Team Memory; Role-based access. Enterprise Custom Quote On-premise hosting; SOC2 compliance; Dedicated memory clusters. 🎁 How to Get Started: Visit We recommend the "Memory Stress Test": upload a 50-page technical manual and a week of chat logs, then ask your agent a question that requires connecting a detail from the manual to a specific comment in the chat.