Blue Note Memory MCP-native memory

Early access · founding pricing

Simple pricing. Start free.

Every plan is MCP-native and includes the full memory stack — persistent, scratch, episodic, graph, and the governed learning loop — plus per-phase model routing (harness), governed skills, and read-only custom tools. Tiers differ by size, retention, skills/tools limits, and support. 14-day free trial, cancel anytime.

Monthly Annual 2 months free

Developer

For one builder wiring memory into agents.

$19 / mo

  • 2 users · 5 agent tokens
  • 35,000 memories (~500 MB)
  • 150,000 calls / month
  • Persistent · scratch · episodic · graph · learning loop
  • Harness model-routing · 10 skills · 5 custom tools
  • 30-day episodic retention
  • Community support
Most popular

Team

For small teams sharing agent memory.

$99 / mo

  • 8 users · 20 agent tokens
  • 350,000 memories (~5 GB)
  • 2,000,000 calls / month
  • Team, project & client scopes
  • Harness model-routing · 50 skills · 25 custom tools
  • 12-month episodic retention
  • Email support (next business day)

Business

For teams scaling agent memory across an org.

$299 / mo

  • 25 users · 65 agent tokens
  • 1,500,000 memories (~20 GB)
  • 7,500,000 calls / month
  • Tenant & admin canon scopes + audit
  • Harness model-routing · 200 skills · 100 custom tools
  • 24-month episodic retention
  • Priority support

No charge during your trial. Cancel anytime from the billing portal. Prices in USD.

Need local custody or a private deployment?

Hybrid and on-prem run Blue Note Memory on your infrastructure — your data never leaves your boundary. Priced to your footprint.

Hybrid

Your agents keep memory local via the appliance, with a hosted control plane for coordination, updates, and support.

Contact sales

On-Prem (Self-Install or Managed)

Run the full appliance in your own environment — self-installed under a commercial license, or deployed and operated by us on your infra.

Contact sales