Multi-Agent

A team of researchers on one question

A team of research agents each take a different angle on your question, search the web, then synthesize one structured answer. Run it on a standalone question, or generate a list.

Try it for freeGive to your AI

▶ 2-min demo video coming soon

Diagram showing one research question fanning out to several research agents, each on a different angle, then converging into one synthesized answer
$0.30 to $2per question
3 to 6 agentsper question, by effort level
🔬

1 question → 6 agents → 1 answer

Research a Hard Question

Survey the state of formal verification for AI systems across six research angles, synthesized into one structured brief.

$1.43 • high effort • 6-field answer

📇

1 prompt → 52-row table

Generate a List

Find startups selling data, benchmarks, and RL environments to frontier AI labs. One prompt returns a full table.

$1.07 • 52 companies • 8 categories

⚖️

compare → structured summary

Compare Options

Compare the major prediction-market platforms on mechanism, volume, and regulatory status. A how-to guide.

$0.36 • low effort • ~4 minutes

Give your AI a research team

Claude Code

claude mcp add futuresearch --scope project --transport http https://mcp.futuresearch.ai/mcp

Then ask Claude to research a question.

Python SDK

pip install futuresearch

import pandas as pd
from futuresearch.ops import multi_agent
result = await multi_agent(
  task="Research the state of
    formal verification for AI",
  input=pd.DataFrame(),
)

Pricing

Start with $20 in free credits. No credit card required. You pay for the LLM tokens each run uses. The figures below are from real production runs.

QuestionEffortOutputCost
Formal verification for AIHigh6-field brief$1.43
Startups selling to AI labsHigh52-row table$1.07
Prediction-market platformsLowComparison summary$0.36

Why costs vary

Cost scales with the effort level. LOW puts 3 agents on the question, MEDIUM 4, and HIGH 6. More agents cover more angles and read more pages. A standalone question runs about $0.30 at low effort and up to $2 at high effort.

Set return_list to turn one question into a table, with one row per item found.

Resources

Guides

SDK Docs

Case Studies

Put a research team on your question

Start with $20 free credit. No credit card required.

Try it for freeSee pricing