tierproxy vs the field

Capability	tierproxy	Smartproxy	Bright Data	Oxylabs	DataImpulse
Multi-provider aggregation	✅ native	single-provider	single-provider	single-provider	single-provider
Client-side smart routing (cheapest/fastest/most_reliable)	✅ built-in	manual	manual	manual	manual
Live SSE usage stream	✅	❌	❌	❌	❌
MCP server (Claude/Cursor/Cline)	✅	❌	❌	❌	❌
Pre-built AI/ML framework examples	10	0	1 (LangChain)	0	0
OpenAPI 3.1 spec	✅	❌	❌	❌	❌
OpenTelemetry built-in	✅ opt-in extra	❌	❌	❌	❌
Pip-installable CLI	✅ tierproxy	❌	❌	❌	❌
License	Apache-2.0 (this SDK)	proprietary	proprietary	proprietary	proprietary
Sync + async parity	✅	partial	partial	partial	partial
Built-in budget guardrails	✅	❌	❌	❌	❌

When to pick which

• You’re an AI/ML engineer scraping for RAG → tierproxy. The cookbook page has 5 ready recipes for the popular frameworks; no other vendor ships these.

• You’re locked into a single provider’s contract → use that vendor’s SDK directly; tierproxy is built for multi-provider aggregation.

• You need browser automation with proxy → tierproxy + Playwright via playwright_proxy_config (see cookbook recipe 3).

• You’re cost-sensitive → tierproxy’s routing="cheapest" is unique; no competitor selects upstream by price per request.

• You need observability → only tierproxy ships OpenTelemetry today.