Security &
Trust.
We design Remii from the ground up to protect your integrations, maintain complete privacy, and run resilient virtual workflows.
Sandboxed Execution
All code runs inside secure, isolated sandboxes. This prevents external tools from modifying files, compromising system resources, or launching unauthorized network calls.
Data Isolation
Your chat histories, profile preferences, and memories are stored in a dedicated database with strict row-level segregation, ensuring your context is never leaked or shared.
Secure OAuth Integrations
Remii connects to your productivity apps (GitHub, Slack, Google Calendar) using secure OAuth 2.0. Access tokens are fully encrypted and only transmitted during tool invocation.
State Machine Resilience
Built-in watchdog reconcilers monitor active agent runs. If a tool call times out or encounters network degradation, the state machine restarts and recovers dynamically without losing context.
Privacy First Design
No global data profiling. We do not inspect your prompts, cache files, or memory indexes. You retain complete authority to delete or clear data instantly.
SSL/TLS Transit
All communications between the browser, tRPC endpoints, database, and Redis cache are fully encrypted in transit using industry-standard TLS 1.3.
Have questions?
Our code is fully transparent, and we are committed to making your assistant secure. If you have identified a vulnerability or have infrastructure-specific questions, please reach out to us.