Tenure

Tenure is an AI memory and context-management tool that gives coding assistants a persistent, auditable knowledge layer for solo developers and engineering teams. Instead of relying on vector search or a single context window, Tenure captures "structured beliefs" with provenance, versioning, and scope boundaries, then injects the right context at generation time to prevent AI mistakes before a pull request is even opened. It ships with four memory modes (observe-only, curated, adaptive, and document-driven) and an observe-before-inject safety model, so teams control exactly what the AI learns and remembers. Tenure works across VS Code, VSCodium, Windsurf, Continue, Open WebUI, OpenClaw, and LibreChat, and supports OpenAI-compatible BYOK endpoints. It advertises source-backed provenance, sub-15ms retrieval latency, 1.0 retrieval precision, and a 0.00 drift score, plus features like Project Resume for restoring context across repositories and persistent knowledge for PR review. Tenure is MIT-licensed and open source, installable via Docker in about 30 seconds, with Helm chart and OIDC/SCIM support for larger deployments. Pricing is not published on the site; the tool is offered as free, self-hosted open source.

Solo developers and engineering teams use Tenure to give AI coding assistants persistent, auditable, cross-session memory and controlled context injection.