Pricing

Start free.
Pay for what you use.

Usage-based, not seat-based. Cache hits cost 1 unit. LLM calls cost 3–5. Scale without a sales call.

Trial

$0

200 units to get started

  • 200 units included
  • 1 project
  • All 3 endpoints
  • Community support
Start free

Pay as you go

$0.80

per 1,000 units

  • After trial
  • Unlimited projects
  • All 3 endpoints
  • Email support
  • Usage dashboard
Get API key

Enterprise

Custom

SLA + dedicated support

  • Volume pricing
  • Unlimited projects
  • On-prem option
  • Dedicated support
  • Custom SLA
Contact us

How units are counted

Not all calls cost the same.

Cache hits are cheap. LLM calls cost more. You're billed for what actually runs.

/context — cache hit
units:1
note:served from cache
/context — cache miss
units:3
note:LLM call required
/context — no-cache
units:5
note:always re-runs LLM
/remember
units:3
note:extract + embed + store

FAQ

One compute unit = one minimal cache hit. Heavier operations (cache miss, fresh mode, remember) cost 3–5 units because they involve LLM calls or vector writes.

Requests are paused until you add a payment method. After that you pay $0.80 per 1,000 units with no hard cutoff.

No. You get 200 trial units with no card required. Add one only when you want to continue beyond the trial.

Yes. Set a monthly unit cap in your dashboard settings. You can choose to receive email alerts at a custom threshold, or enable a hard stop that blocks requests once the cap is reached.