Pay for memory, not anxiety

Start free. Upgrade when your agents get serious. Every plan includes compaction, priority eviction, and observability.

Monthly Annual −17%

Starter

For side projects and experiments.

$0 /mo

free forever

  • 1M managed tokens / month
  • 1 project, 3 agents
  • Smart compaction + priority eviction
  • Community support
Start free

Team

For orgs running fleets of agents.

$99 / seat / mo

per seat / month, billed monthly

  • 500M managed tokens / month
  • Multi-agent shared context pools
  • SSO / SAML, audit logs
  • Data residency options
  • 99.9% uptime SLA
  • Dedicated success engineer
Talk to sales

Compare plans

CapabilityStarterProTeam
Managed tokens / month1M50M500M
Smart compaction
Priority-based eviction
Semantic recall
Custom compaction policies
Multi-agent shared context
Observability dashboardBasicFullFull + audit
SSO / SAML
Uptime SLA99.9%
SupportCommunityPriorityDedicated

Frequently asked questions

Any token that passes through the compaction engine — messages stored, compacted, evicted, or recalled. Raw passthrough without management doesn't count against your quota.
Yes. Context Controller sits in front of any OpenAI-compatible API, and our Python and JavaScript SDKs wrap the major providers directly. If your provider speaks the chat-completions shape, it works.
Nothing breaks. Compaction and recall keep working; overage is billed at a flat per-million-token rate shown on your dashboard, and you can set hard caps per project if you prefer.
Never. Your prompts, tool outputs, and recalled content are yours alone. We don't train models on customer data, and Team plans can pin data residency to a region.
Team plans include a self-hosted option (single binary + Postgres) for regulated environments. Contact us for the deployment guide.

Still deciding? Run the demo.

See exactly what your money buys — the compaction policy, live in your browser.