Brime
Search, extract, and research the web with one API key.
Set up Brime via MCP
Free. No account, no API key needed for first use. Works in every major coding agent.
Claude Code
Run in terminal:
claude mcp add brime -- npx -y mcp-server-brimeCodex
Run in terminal:
codex mcp add brime -- npx -y mcp-server-brimeCursor
One-click install →
VS Code
One-click install →
No config needed up front — the first time your agent calls a Brime tool, it asks for your API key and saves it locally. Get a free key at brime.dev/api-keys (1,000 credits/month, no card). See the MCP guide for advanced config and the brime_set_key tool reference.
Build with Brime APIs
Pick a product, copy the setup prompt, and paste it into your coding agent — or use the SDK directly.
Fresh, ranked excerpts for grounding answers.
from brime import Brime
client = Brime(api_key="sk-brime-...")
result = client.search(query="Apple Q4 2025 earnings highlights")
print(result.answer)import { Brime } from "@brime-ai/brime";
const client = new Brime({ apiKey: "sk-brime-..." });
const result = await client.search({ query: "Apple Q4 2025 earnings highlights" });
console.log(result.answer);curl https://api.brime.dev/v1/search \
-H "Authorization: Bearer sk-brime-..." \
-H "Content-Type: application/json" \
-d '{"query": "Apple Q4 2025 earnings highlights"}'Clean markdown from any URL. PDF, DOCX, HTML, SPA — all unified.
result = client.extract(urls=["https://example.com"])
for r in result.results:
print(r.url, r.method, len(r.markdown))const result = await client.extract({ urls: ["https://example.com"] });
for (const r of result.results) {
console.log(r.url, r.method, r.markdown.length);
}curl https://api.brime.dev/v1/extract \
-H "Authorization: Bearer sk-brime-..." \
-H "Idempotency-Key: $(uuidgen)" \
-H "Content-Type: application/json" \
-d '{"urls": ["https://example.com"]}'Multi-step research with citations. Single tuned flow, ~30–90 s sync.
result = client.research(query="compare frontier coding models")
print(result.answer)const result = await client.research({ query: "compare frontier coding models" });
console.log(result.answer);curl https://api.brime.dev/v1/research \
-H "Authorization: Bearer sk-brime-..." \
-H "Idempotency-Key: $(uuidgen)" \
-H "Content-Type: application/json" \
-d '{"query": "..."}'What's in the box
Search
Instant SERP snippets plus a short synthesised answer — sub-second.
Extract
PDF, DOCX, HTML, SPA — everything to clean markdown. Auto Chrome escalation.
Research
Iterative agent with citations. Single tuned flow, ~30–90 s, optional SSE stream.
Python SDK
pip install brime — sync + async, fully typed.
Node.js SDK
npm install @brime-ai/brime — ESM + CJS, native fetch, zero deps.
LangChain
langchain-brime — 3 tools + retriever, drop-in for any chain.
Drop-in for other vendors
Already using Tavily, Exa, or Parallel? Point your existing client at Brime — no code changes: