Compare · Web Search
Brave Search vs Perplexity API Platform MCP Server
Brave Search exposes 8 tools and adds about 8,289 tokens to every model turn. Perplexity API Platform exposes 4 tools and adds about 1,420. Brave Search costs 5.8× as much context. Measured by connecting to each server and counting its tool list, not from the README.
 Brave Searchbrave |  Perplexity API Platformperplexityai | |
|---|---|---|
| Context tax | ~8,289 tok/turn | ~1,420 tok/turn(leaner) |
| Tools | 8 | 4 |
| Cost per 1,000 turns (Claude Sonnet 5.5, uncached) | $16.58 | $2.84 |
| Security score | 97/100 | 100/100 |
| Cold start | 5,865 ms | 732 ms |
| Runs | Local or hosted | Local (stdio) |
| Needs keys | BRAVE_API_KEY | None declared |
| Freshness | Active | Active |
| GitHub stars | 1.5k | 2.6k |
| License | MIT | MIT |
| Official MCP registry | Yes | Yes |
Install in Claude Code
Brave Search
claude mcp add brave-search-brave -e BRAVE_API_KEY='<brave-api-key>' -- npx -y @brave/brave-search-mcp-server
Perplexity API Platform
claude mcp add perplexity-ai -- npx -y @perplexity-ai/mcp-server
Tools
By name, from each server's live tool list. The same name can still behave differently.
Only in Brave Search · 8
- brave_web_search
- brave_local_search
- brave_video_search
- brave_image_search
- brave_news_search
- brave_summarizer
- brave_llm_context
- brave_place_search
In both · 0
- none
Only in Perplexity API Platform · 4
- perplexity_ask
- perplexity_research
- perplexity_reason
- perplexity_search
Questions
- Which uses fewer tokens, Brave Search or Perplexity API Platform?
- Perplexity API Platform: its tool definitions take about 1,420 tokens per turn against 8,289 for Brave Search (5.8× more). Tool definitions are resent on every model turn.
- Do Brave Search and Perplexity API Platform have the same tools?
- No tool names overlap: Brave Search has 8 tools and Perplexity API Platform has 4.
- Are Brave Search and Perplexity API Platform hosted or local?
- Brave Search: local or hosted. Perplexity API Platform: local (stdio).
Token counts use the o200k tokenizer on each server's tool definitions as an agent receives them; clients that load tools lazily pay less. Prices are Claude Sonnet 5.5 input at $2/MTok. Want to check one yourself? Run npx mcpskillhub test <package> — how it works.