Features
Walkthroughs display diffs in logical order
Refactoring agent auto-fixes duplicates, secrets, long functions
Per-line feedback to AI agents without context pollution
Image attachments in agent feedback
Parallel workspace management for multiple projects
Custom walkthroughs beyond diff scope (e.g., bug explanation)
Configurable refactoring explanation detail (thorough/standard/terse) with code diff
Automatic workspace archiving and re-inflation to save disk space
Integrated agent terminal with TUI support
Credit usage alerts via toast notifications
Hotkey suggestion toasts with auto-dismiss
Sleep mode control to keep agents running when laptop is closed
Bulk feedback writing across multiple comments
Walkthrough over any diff range (June 2026)
Claude 5 Fable support for walkthroughs and refactorings (June 2026)
Trained exclusively on production code for maintainability
Supports COBOL, Fortran, Verilog, Rust, SQL
Terminal-native CLI with local-to-remote execution
External tool access via MCP
Multi-agent orchestration via Cosine Swarm
Cosine Cloud for parallel tasks and shared projects
Scout model: on-device, post-trained from Devstral 123B
Outpost model: 59.3% on Niche-Bench, post-trained from Kimi K2.6
Frontier model: coming soon for frontier-scale reasoning
Air-gapped deployment option
Managed single-tenant cloud
Credit-based usage metering
CLI built-in search (v1.16)
Lumen Sovereign coalition for sovereign AI
