API v1 is live

Ground every AI answer in verified, cited sources

Groundlink is a pay-per-query API that AI models call on every search. It returns verified results — real URLs, snippets, and a source label — so chatbots, agents, and assistants answer with citations instead of hallucinating.

Read the API docs Buy credits How it works

POST https://9ea69cec60fa01f65bbb647a092bcbb4.ctonew.app/api/v1/ground · Bearer auth · typically <1s

How it works

One request in, cited results out. The response is small, flat, and shaped for LLMs: every result carries the fields a model needs to cite its answer.

The model sends a query

Your agent, chatbot, or assistant calls Groundlink with a natural-language question — automatically, on every search where it needs a verifiable answer.

Groundlink returns verified results

Free, reliable sources (Wikipedia and DuckDuckGo for v1) return top matches with a real URL, a readable snippet, and a source label — no invented facts.

The model answers with citations

Every claim in the reply can point at a URL the user can click. The answer is correct because the model grounded it in what the sources actually say.

What the model actually gets

✓title — a clean, human-readable heading
✓url — a real, clickable source link
✓snippet — a short excerpt the model can quote
✓source — a label for attribution ("wikipedia", "duckduckgo")

{
  "query": "James Webb Space Telescope",
  "results": [
    {
      "title": "James Webb Space Telescope",
      "url": "https://en.wikipedia.org/wiki/James_Webb_Space_Telescope",
      "snippet": "The James Webb Space Telescope (JWST) is a space
telescope designed to conduct infrared astronomy...",
      "source": "wikipedia"
    },
    {
      "title": "James Webb Space Telescope Category",
      "url": "https://duckduckgo.com/c/James_Webb_Space_Telescope",
      "snippet": "James Webb Space Telescope Category",
      "source": "duckduckgo"
    }
  ],
  "meta": {
    "request_id": "7b36dfd4-3ca3-408c-9791-96c1bf2e4603",
    "result_count": 2,
    "processing_ms": 500
  }
}

Built for AI systems that must be right

If your product generates language, Groundlink plugs into its tool-calling loop — one HTTPS call, no SDK required.

Chatbots

Product-facing assistants that need every answer traceable to a source users can open.

AI agents

Autonomous agents that browse, summarize, and act — grounding keeps their decisions on real facts.

Copilots & assistants

Developer tools and workplace copilots that must return verifiable, professional answers.

Model platforms

API and platform providers that want retrieval-grounded outputs by default, at scale.

Pricing that scales with volume

pay-per-query micro-fees — from under one cent per query — so grounding is an automatic step in every search, not a budget decision.

Pay-per-query

Core model

Pay-per-query micro-fees: each grounding call costs a micro-fee — cheap enough that an AI can call it automatically on every search, with no human in the loop. Revenue comes from massive transaction volume, not big per-customer fees.

from under one cent per query

Free tier for testing

Get an API key and run real queries against the live endpoint before committing anything — no card required to start.

No charge

Prepaid credits

A human funds the account once; the AI auto-consumes credits per call. Volume tiers kick in as usage scales. This is the billing shape — wiring happens before launch.

Fund once, run forever

Indicative only — final per-call pricing will be set by Groundlink before billing goes live. The pricing model above is the plan; specific per-call figures are decided by Groundlink before billing is wired.

Plugs into any tool-calling loop

Groundlink is a plain HTTPS + JSON API. Advertise ground_search as a tool, then call the endpoint when the model asks. Works with OpenAI-style function calling, Anthropic tool use, MCP, or any framework that can issue an HTTP request.

✓ No SDK to install — one fetch call
✓ CORS enabled; call it from anywhere
✓ Usage metered per API key

See the integration guide

// Advertise the tool to the model (OpenAI-style function calling)
const tools = [{
  type: "function",
  function: {
    name: "ground_search",
    description: "Search Groundlink for verified, citable web results.",
    parameters: {
      type: "object",
      properties: {
        query: { type: "string", description: "The query to ground." },
        max_results: { type: "integer", minimum: 1, maximum: 10 }
      },
      required: ["query"]
    }
  }
}];

// When the model emits a tool call, fetch verified results
const res = await fetch(BASE_URL + "/api/v1/ground", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Authorization": "Bearer glk_..."
  },
  body: JSON.stringify({ query: "James Webb Space Telescope" })
});
const { results } = await res.json();
// -> answer with citations from results[].url

Stop answering from memory. Start answering from sources.

One page of docs, one endpoint, one API key. Get your first grounded response in minutes.

Read the docs — it's one page