Per-token pricing with annual commitments. Open-weight models discounted 10%+ and closed models 5%+ compared to list rates. No platform fees or per-seat charges. Rates improve as committed spend grows.
Key features
AI agents that autonomously plan, write, and review code
Works across desktop IDE, CLI, web and mobile
Understands entire codebase using advanced codebase search
Multiple AI models including Grok, GPT-5, and Opus
Team rules to teach preferences and architectural decisions
Fast codebase search using Instant Grep for large repositories
Cloud agents that work in parallel for hours or days
Git and checkpoint system to track code evolution
Extensible with plugins, skills, and MCP integrations
End-to-end encryption on every connection
Single-tenant dedicated deployments
300+ models through one endpoint
OpenAI-compatible API
Zero data retention enforced at gateway
Agents API for coding agents
SAML SSO and SCIM provisioning
PII removal before closed models
What makes it different
Agents work autonomously, eliminating manual coding tasks
Available everywhere you work across all development tools
Understands full codebase context before making changes
Choose best model for each task with multi-model support