The local stack is free.
Everything that optimizes your sessions runs on your machine and costs nothing — bring your own API key or subscription. Paid plans add the hosted layer: metering, dashboards, team fleet view, and enterprise controls. They are not live yet.
The full local optimization stack, on your machine.
- Local gateway — record and apply, OpenAI + Anthropic shapes
- Hybrid live-history input compaction (on-device local model)
- Task-aware output shaping — gateway and native hooks
- Claude Code, Codex, and Cursor
- Content-free receipts and byte-exact recovery
- Soft local weekly usage caps
Hosted meter and evidence dashboard for individuals.
- Everything in Free / Local
- Hosted usage meter (content-free counts only)
- Evidence dashboard
- Higher caps
Fleet visibility and verifiable receipts for teams.
- Everything in Pro
- Seats and fleet view
- Signed receipt verification
- Metered overage
Controls and deployment options for regulated environments.
- Everything in Team
- SSO / RBAC / audit
- On-prem / VPC deployment
- Support and SLA
Prices for Pro, Team, and Enterprise are indicative — hosted billing is not live, and nothing on this page is a signup for a paid service today. Usage caps meter volume Compaction counts exactly, from content-free receipts (token and request counts — never your content).
We don't price as a share of savings. The savings we report are measured token counts from real sessions, each carrying its evidence label — and on API-key usage, fewer tokens are directly a smaller bill, since the provider prices every token.