Google Scholar API for AI agents (papers JSON, no API key)

When a user asks you to "find papers on X," "what's the citation count for this work," "pull this author's publications," or "ground this claim in peer-reviewed sources," you need structured scholarly results — titles, links, citation counts, authors, publication venues — as JSON, not scraped Scholar HTML and not a SerpAPI / Custom Search key.

Marketplace for AI Agents exposes three machine-payable endpoints for this: GET /api/public/v1/google-scholar turns a query into organic_results, GET /api/public/v1/google-scholar-author turns an author_id into an author record with articles, and the scholar-pack intent chains papers + optional author in one call. The first 3 calls per agent identity are free; after that each single-engine call costs $0.005 USDC on Base, settled over x402 with an X-PAYMENT header.

Endpoint contracts

Optional google-scholar params: hl (language, defaults to en), num (results per page), and page. Optional google-scholar-author / scholar-pack params: hl; scholar-pack also takes an optional author_id. Each single-engine call is billed at $0.005 USDC.

The free tier is 3 successful calls total per agent identity (a hash of IP + User-Agent), shared across every endpoint on the marketplace — it is not 3 per endpoint. Settlement uses the x402 exact scheme on Base mainnet; append ?network=base-sepolia to any endpoint for testnet USDC and sandbox mock data.

• Listing: https://marketplaceforaiagents.com/listings/google-scholar
• Listing: https://marketplaceforaiagents.com/listings/google-scholar-author
• Listing: https://marketplaceforaiagents.com/listings/scholar-pack
• Markdown twins: /agent/listings/google-scholar.md, /agent/listings/google-scholar-author.md, /agent/listings/scholar-pack.md

Minimal calls (free tier first)

Spend the 3 free evaluation calls before signing anything. These run on the free tier (if you still have quota) and return real JSON:

curl -si "https://marketplaceforaiagents.com/api/public/v1/google-scholar?q=large+language+model+agents"

curl -si "https://marketplaceforaiagents.com/api/public/v1/google-scholar?q=retrieval+augmented+generation&hl=en&num=10"

curl -si "https://marketplaceforaiagents.com/api/public/v1/google-scholar-author?author_id=…"

An organic_results item looks like this (illustrative — always trust the live JSON, not this sample):

{
  "position": 1,
  "title": "Example paper title",
  "link": "https://example.edu/paper.pdf",
  "snippet": "…",
  "publication_info": { "summary": "Authors · Venue · Year" },
  "inline_links": { "cited_by": { "total": 128 } }
}

Note: a 400 missing_required_parameters response does not consume a free call, and the discovery endpoints (/.well-known/x402, catalog.json, openapi.json) never consume free tier.

When to use scholar-pack instead of two calls

• google-scholar alone — user only needs papers / rankings / citation counts
• google-scholar-author alone — user already has an author_id
• scholar-pack — user asked for papers plus that author's record
• google-patents — related IP search (separate endpoint) when the ask is patents not papers

Intents are billed per step: the 402 advertises the chain and steps, the price is $0.005 × steps, and one intent invocation consumes one free-tier call. Prefer scholar-pack when the user wants papers + author so you spend one free-tier slot instead of two.

curl -si "https://marketplaceforaiagents.com/api/public/v1/scholar-pack?q=large+language+model+agents"

Full machine-readable specs live in the catalog (https://marketplaceforaiagents.com/api/public/v1/catalog.json) and the OpenAPI document (https://marketplaceforaiagents.com/api/public/v1/openapi.json).

Paying after the free tier (x402 / USDC)

Once your identity has spent its 3 free calls, the same request returns HTTP 402 with an accepts[] payment requirement. Decode it, verify every field against your policy, sign the EIP-3009 USDC authorization, base64 the x402 payload, and resend with X-PAYMENT:

{
  "x402Version": 1,
  "error": "payment_required",
  "accepts": [
    {
      "scheme": "exact",
      "network": "base",
      "maxAmountRequired": "5000",
      "asset": "0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913",
      "payTo": "0x05e82e03753c7bc99fb24d10a876cce53f24b7b7",
      "resource": "https://marketplaceforaiagents.com/api/public/v1/google-scholar?q=large+language+model+agents"
    }
  ]
}

• maxAmountRequired is 5000 atomic units = $0.005 USDC (6 decimals) per single-engine call.
• Budget your spend before signing: see How AI agents budget x402 API calls.
• Test the full payment path with zero real cost: see the Base Sepolia smoke test.
• A paid 200 carries the settlement receipt in X-PAYMENT-RESPONSE — store it for reconciliation.
• Prefer the MCP server (https://marketplaceforaiagents.com/mcp) for discovery; payment is the auth, no OAuth.

Protocol background: x402 protocol explained. Free-tier mechanics: how the x402 free tier and agent identity work.

Agent playbook (Grok Bot / Claude Cowork / Cursor)

• Keep the research intent in q; use num and page to page through results.
• Extract author_id from Scholar results before calling google-scholar-author.
• Use google-scholar for papers only; use scholar-pack when the user wants an author dossier too.
• On 402, settle with x402 USDC — do not ask for a SerpAPI or Google Cloud key.
• Cite title, link, venue, and citation counts from the JSON; never invent papers or citation numbers.
• Allowlist payTo 0x05e82e03753c7bc99fb24d10a876cce53f24b7b7 and USDC 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913 before signing.

Related entrypoints

• Google Scholar listing: https://marketplaceforaiagents.com/listings/google-scholar
• Google Scholar Author listing: https://marketplaceforaiagents.com/listings/google-scholar-author
• Scholar Pack listing: https://marketplaceforaiagents.com/listings/scholar-pack
• Markdown twins: /agent/listings/google-scholar.md, /agent/listings/google-scholar-author.md, /agent/listings/scholar-pack.md
• x402 manifest: https://marketplaceforaiagents.com/.well-known/x402
• Catalog: https://marketplaceforaiagents.com/api/public/v1/catalog.json
• OpenAPI: https://marketplaceforaiagents.com/api/public/v1/openapi.json
• MCP: https://marketplaceforaiagents.com/mcp
• llms.txt: https://marketplaceforaiagents.com/llms.txt
• Marketplace: https://marketplaceforaiagents.com/marketplace
• Docs: https://marketplaceforaiagents.com/docs
• MCP + x402 guide: https://marketplaceforaiagents.com/learn/mcp-x402-marketplace
• How agents discover APIs: https://marketplaceforaiagents.com/learn/how-ai-agents-discover-apis
• Google Search for agents: https://marketplaceforaiagents.com/learn/google-search-api-for-agents
• Google Maps for agents: https://marketplaceforaiagents.com/learn/google-maps-api-for-agents

Capability first, payment second — find the Scholar endpoint, prove it with free calls, then settle unpaid traffic with x402 USDC on Base.