Reference
Pricing.
Consumer tiers
| Tier | Price | Memories | Connected apps |
|---|---|---|---|
| Free | $0 | 500 | 3 |
| Plus | $9/mo | 10,000 | ∞ |
| Pro | $19/mo | 100,000 | ∞ |
Developer tiers
| Tier | Price | API calls / mo | Hosted tokens | Connected users |
|---|---|---|---|---|
| Sandbox | $0 | 10k | 100k | 50 |
| Starter | $29/mo | 200k | 2M | 1,000 |
| Growth | $199/mo | 2M | 20M | 25,000 |
| Pro | $999/mo | metered | metered | metered |
| Enterprise | custom | — | — | — |
What you're billed for
- API calls — any request to the platform.
- Hosted tokens — input + output tokens through the gateway. Tracked separately from API calls so cost is predictable.
- Memory writes — counted; per-tier soft caps.
- Context requests — vector + graph retrievals.
- Connected users — distinct users with an active connection to your app, billed at end-of-cycle.
BYOK
Bring your own LLM provider key (Anthropic, in current build). PAMbase bills only for identity, memory, context, and graph/vector retrieval — no hosted-token charge. There's a flat platform fee per active connected user (announced at GA).
Marketplace revenue share
Apps listed in the public marketplace get distribution and the standard PAMbase revenue share applies. Details at GA.
Numbers in this build are placeholders. The metering plumbing (usage_counters + per-call increments) is wired and works; billing runs on Paddle (Merchant of Record).