Ask a question. Your agent comes back with the answer, the source and the papers behind it.

A research agent needs three things: a search engine, a way to read what it finds, and the literature. Strale gives it all three through the connection it already has, paid per call.

  • Calls3 tools
  • One run€0.13
  • RoutesAPI, MCP, SDKs and x402

How your agent does it

3 calls, one connection

Someone asks the agent what the research says about large language models in education.

Every step goes through the same key and the same connection, with no account at any of the providers behind them.

  1. Step 1: Google Search €0.10

    Searches Google and gets ranked results, each with its title, link and snippet.

    x402 google-search
  2. Step 2: URL to Markdown Free

    Opens the best result and turns the page into clean text the model can read.

    Free to try x402 url-to-markdown
  3. Step 3: Academic Paper Search €0.03

    Finds the papers behind the claims, with authors, citations and an open-access link.

    x402 academic-paper-search
  4. The result

    The agent answers from a page it actually read, and cites a paper it can link to.

Illustrative example

One run, start to finish

What the agent sends and what comes back, step by step.

  1. 1 Google Search

    Agent sends
    {
      "query": "large language models in education research"
    }
    Comes back
    {
      "query": "large language models in education research",
      "results": [
        {
          "position": 1,
          "title": "Large language models in the classroom: a review",
          "url": "https://example.org/llm-education-review",
          "snippet": "A review of how schools and universities use large language models, and what the studies found."
        }
      ]
    }
  2. 2 URL to Markdown

    Agent sends
    {
      "url": "https://example.org/llm-education-review"
    }
    Comes back
    {
      "title": "Large language models in the classroom: a review",
      "markdown": "# Large language models in the classroom\n\nStudies since 2023 point to three uses that hold up…",
      "word_count": 2140
    }
  3. 3 Academic Paper Search

    Agent sends
    {
      "query": "large language models",
      "limit": 1
    }
    Comes back · published example
    {
      "papers": [
        {
          "doi": "https://doi.org/10.1016/j.lindif.2023.102274",
          "type": "article",
          "title": "ChatGPT for good? On opportunities and challenges of large language models for education",
          "venue": "Learning and Individual Differences",
          "authors": [
            "Enkelejda Kasneci"
          ],
          "abstract": null,
          "openalex_id": "https://openalex.org/W4323655724",
          "cited_by_count": 6326,
          "is_open_access": true,
          "open_access_url": "https://epub.ub.uni-muenchen.de/125071/1/ChatGPT_for_Good_v3.pdf",
          "landing_page_url": "https://doi.org/10.1016/j.lindif.2023.102274",
          "publication_date": "2023-03-09",
          "publication_year": 2023
        }
      ],
      "returned": 1
    }

What one run costs

Priced per call from the live catalogue, with no subscription and no provider accounts.

One run, with an API key €0.13

Free tools cost nothing when called with a key. Each tool is charged only when its call succeeds.

No account at all

Pay per request from a wallet

Every step of this run is eligible for x402. Your agent pays each request in USDC as it goes: $0.1944 for the run, quoted before anything is paid.

The calls, in code

Each step through the REST API, with one key and the example’s inputs. Your agent passes what one step returns into the next.

The same tools work over MCP and the SDKs, and keyless over x402. Each tool’s page has the call for every route.

Read the docs
# Needs an API key. max_price_cents caps what this call may cost.
curl -X POST https://api.strale.io/v1/do \
  -H "Authorization: Bearer $STRALE_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{
    "capability_slug": "google-search",
    "inputs": {
      "query": "large language models in education research"
    },
    "max_price_cents": 10
  }'

Give your agent this job today

Start with Google Search, then add the next step through the same connection.