See what the best
agents do differently.

import { observe } from "@virajmishra1/bench-sdk";

const agent = observe({
  apiKey: process.env.BENCH_KEY,
  agent: "mcpify",
  model: "claude-sonnet-4",
  framework: "vercel-ai",
  tools: ["web-search", "calculator"],
});

await agent.task("search", { query }, async (t) => {
  const result = await doSearch(query);
  t.log("found_chunks", result.length);
  t.cost(0.004);
  return result;
});

from bench import observe

agent = observe(
    api_key=os.environ["BENCH_KEY"],
    agent="mcpify",
    model="claude-sonnet-4",
    framework="langchain",
    tools=["web-search", "calculator"],
)

result = await agent.task("search", {"query": query}, search_fn)

async def search_fn(t):
    result = await do_search(query)
    t.log("found_chunks", len(result))
    t.cost(0.004)
    return result