Every AI interaction you initiate is billed transparently. Every background job the system runs for you is free. You pay for what you ask — never for what the system does silently to keep your world running.
Included Lift comes with your plan. Prepaid top-ups never expire. Metered overage kicks in only when you need more — and you control the budget.
Every plan includes Lift Power per billing period. Free gets 100K. Pro gets 2M. Business gets 5M. Scientist gets 50M. It resets automatically. You never lose it — it just refills.
Run out of included Lift? Buy 250K more. Top-ups never expire — they sit in your account until you use them. Available on every plan, including Free.
On paid plans, keep going past your quota with metered billing. Set a monthly budget. Get warnings before you hit it. Enable a hard stop — or let it flow. Your choice, your control.
Embedding sync, pattern scanning, scheduled focus jobs, daily trend summaries, memory compression, semantic routing — these run in the background as infrastructure. They never deduct from your Lift. They never appear on your bill. They are the engine that keeps your world alive, and they are included in your subscription.
Always free — never billed:
When content changes, only the changed blocks are re-processed. Bloom deduplication skips 30% of duplicates automatically. The result: updating 100 pages in a 1M-page corpus costs about $2 and takes 1.5 minutes. Others re-ingest everything.
100 pages changed
~1.5 min · $2
Full re-ingest · Full cost
Bloom dedup skip
30% fewer LLM calls
0% — re-process everything
Cache hit (warm query)
~50ms · $0
Full pipeline · Full cost
Cross-tenant cache
2nd user: $0 (7-day TTL, 14-day on Lab)
Not available