Tenure is an open-source AI memory layer designed to provide persistent, governable, and scoped state for AI systems. It ensures that AI models have reliable, structured knowledge with clear origins, versions, and scopes, rather than relying on temporary context or fuzzy search methods. This approach allows teams to build AI assistants that accurately remember decisions without mixing data from different projects.
Key Features and Functionality:
- Project Resume: Enables users to pick up where they left off in any repository, ensuring continuity across projects.
- Scoped Tokens: Limits what each client or agent can read, write, extract, and inject, providing controlled access to information.
- Memory Modes: Offers adaptable memory settings to match a team's trust level, including observe-only, curated, adaptive, or document-driven memory.
- GroundEval: Transforms observed agent behavior into reviewable policies, enhancing governance and compliance.
- Editor Integrations: Supports native AI provider integration with editors like VS Code and VSCodium, facilitating seamless development workflows.
- Compatible Clients: Integrates with various clients, including mobile platforms, Open WebUI, and OpenClaw, ensuring broad accessibility.
Primary Value and User Solutions:
Tenure addresses the challenge of memory drift in AI systems, where old or irrelevant information can mix with new data, leading to incorrect outputs. By providing a structured memory system with clear provenance and scope, Tenure eliminates this issue, ensuring that AI models deliver accurate and reliable results. This is particularly beneficial for engineering teams, as it allows for consistent code reviews, shared knowledge across teams, and compliance with governance standards such as the EU AI Act. Additionally, Tenure's integration with popular development tools and its adaptable memory modes make it a versatile solution for various AI applications.
