Hard facts side by side: free tiers, prices, limits, SDKs and privacy. Every value links to the official page it came from and is re-checked every week.
| Fact |  Perplexity API |
|---|---|
| Starting price | $1 per 1,000 fast web searches Pay as you go, no subscription. Model tokens extra, e.g. openai/gpt-6-luna $0.10 in, $0.50 out per 1M. Source, Nov 8, 2026 |
| Rate limit | 1 request/s at Tier 0 Agent API, shared across keys. Tier 1 ($50+ credits bought) gets 3 QPS; higher tiers get more. Source, Nov 8, 2026 |
| Official SDKs | Python |
Official SDKs for Python 3.10+ and Node.js 20+ (TypeScript).
| Trains on your data | No API Terms say Customer Content is not used to train models; zero data retention is not the default. Source, Nov 8, 2026 |
|---|
| EU hosting | No Trust Center lists all 23 subprocessors, incl. AWS, Google Cloud and Azure, in the US; no EU region offered. Source, Nov 6, 2026 |
|---|
| Compliance | SOC 2 SOC 2 Type II report; a 2025 HIPAA gap assessment is listed but is not a certification. Source, Nov 8, 2026 |
|---|
Wirestack