Code privacy by default
Private project files are not used for Vyre product training. AI requests should only include the project context a developer chooses or that the workflow explicitly requires.
Trust & Security
Security by Design
We treat your codebase as sensitive by default. VYRE IDE is being hardened in beta with practical controls, transparent model context, and a clear compliance roadmap.
Compliance Roadmap
Formal compliance certification and deeper enterprise controls are planned, tracked, and not advertised as complete during beta.
Encrypted Transport
Hosted endpoints use encrypted transport, and collaboration should expose whether a session is direct or relayed.
Context Control
Ignored files and private project context should stay out of AI prompts unless the user explicitly includes them.
Authentication
VYRE uses OAuth-based sign-in where providers are configured. We never ask for or store your passwords. Hosted sessions are managed with short-lived tokens stored in HttpOnly, SameSite cookies to reduce XSS and CSRF exposure.
Code Privacy & AI Agents
Private project files are not used for Vyre product training. When you use VYRE Agentic AI, selected context and prompts may be sent to the configured model provider needed to answer the request.
Collaboration & Networking
Vyre Collab uses WebRTC to establish direct Peer-to-Peer (P2P) connections whenever possible. If direct connection is blocked by restrictive firewalls, traffic can fall back to relay infrastructure over encrypted transport.
Security commitments during beta
Reviewable AI activity
Agentic work should remain inspectable through plans, diffs, checkpoints, terminal evidence, and human approval before broad changes are merged.
Account and session hardening
OAuth sign-in, HttpOnly session cookies, SameSite cookie behavior, and short-lived hosted sessions reduce common account and browser attack surfaces.
Collaboration transparency
Live sessions should make connection mode, participant roles, and relay fallback visible so teams know how code and collaboration traffic is moving.
Security FAQ
Does Vyre use private code for product training?
Vyre's public security position is that private project files are not used for Vyre product training. Selected context may be sent to configured model providers when needed to answer an AI request.
How should teams evaluate Vyre security during beta?
Teams should review OAuth setup, session behavior, ignored-file boundaries, AI context controls, collaboration permissions, relay behavior, and Git review checkpoints on a non-critical pilot repository first.
Is Vyre claiming completed enterprise compliance?
No. Vyre describes formal compliance certification as a roadmap item and avoids advertising enterprise certifications as complete during private beta.