# API keys

All keys are optional. Put them in `.env` (never committed); `scripts/configure.sh` writes them into the `env` block of `~/.openclaw/openclaw.json` and into each MCP server's `env`. Servers whose required key is missing are skipped.

| Key | Used by | Free tier | Sign up |
|-----|---------|-----------|---------|
| `BRAVE_API_KEY` | OpenClaw built-in `web_search`, search-layer, last30days | 2,000 q/mo | https://brave.com/search/api/ |
| `EXA_API_KEY` | exa MCP (semantic search, deep researcher, company/people) | yes | https://exa.ai |
| `FIRECRAWL_API_KEY` | firecrawl MCP (JS-rendered scraping) | 500 credits/mo | https://firecrawl.dev |
| `TAVILY_API_KEY` | search-layer skill | 1,000 q/mo | https://tavily.com |
| `PERPLEXITY_API_KEY` | search-layer skill | paid | https://perplexity.ai/settings/api |
| `SEMANTIC_SCHOLAR_API_KEY` | paper-search MCP (1 to 10 RPS) | yes | https://www.semanticscholar.org/product/api |
| `NCBI_API_KEY` | paper-search MCP, pubmed-edirect (3 to 10 RPS) | yes | https://www.ncbi.nlm.nih.gov/account/ |
| `CORE_API_KEY` | paper-search MCP | yes | https://core.ac.uk/services/api |
| `SPRINGER_API_KEY` | paper-search MCP | yes | https://dev.springernature.com |
| `IEEE_API_KEY` | ieee-xplore MCP (~200 calls/day) | yes, manual approval | https://developer.ieee.org/member/register |
| `SCRAPECREATORS_API_KEY` | last30days (Reddit, TikTok, Instagram) | trial | https://scrapecreators.com |
| `XAI_API_KEY` | last30days X backend, search-layer Grok | paid | https://console.x.ai |
| `OPENAI_API_KEY` | last30days enrichment (optional) | paid | https://platform.openai.com |

No key needed: arXiv, bioRxiv, medRxiv, Google Scholar (via paper-search), BCC Research, fetch, knowledge-graph-memory, Wikipedia.

Deferred, not integrated: Google Scholar via SerpAPI (paid), Web of Science Starter API, EBSCO Business Source Complete (institutional credentials).