Strut AI Grounding Engine — Developer Integration

curl -X POST "$STRUT_BASE_URL/v1/ground" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $STRUT_API_KEY" \
  -d '{
    "query": "What are topological quantum qubits?",
    "max_sources": 3
  }'

curl -X POST "https://your-deployment.example.com/v1/search" \
  -H "Authorization: Bearer strut_live_..." \
  -H "Content-Type: application/json" \
  -d '{"query": "who invented the transistor"}'

{
  "query": "What are topological qubits?",
  "grounded_answer": "Topological qubits store quantum information in non-local topological properties of matter [1]...",
  "citations": [
    {
      "source_index": 1,
      "title": "Topological quantum computer - Wikipedia",
      "url": "https://en.wikipedia.org/wiki/Topological_quantum_computer",
      "verified_quote": "Topological qubits are quantum bits that store information in non-local...",
      "relevance_score": 0.8842
    }
  ],
  "model": "strut/web-grounding",
  "telemetry": {
    "fetchMs": 450,
    "extractMs": 18,
    "chunkMs": 2,
    "rerankMs": 6,
    "synthesis_ms": 980,
    "total_end_to_end_ms": 1456
  }
}

{
  "engine": "Strut Grounding Engine",
  "pricing": { "cost_per_query": 0.007, "currency": "USD" },
  "providers": { "web_search_index": { "mojeek_configured": false } }
}

curl -N -X POST "$STRUT_BASE_URL/v1/ground" \
  -H "Authorization: Bearer $STRUT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "...", "stream": true, "include_answer": true}'

from langchain.tools import tool
import requests

@tool
def strut_grounding_tool(query: str) -> str:
    """Search the live web and obtain factually grounded synthesis with verified citations."""
    res = requests.post(
        "$STRUT_BASE_URL/v1/ground",
        json={"query": query, "max_sources": 4},
        headers={"Content-Type": "application/json"}
    )
    if not res.ok:
        return f"Lookup error: {res.text}"
    data = res.json()
    return data.get("grounded_answer", "")

# Attach to LangChain Agent
from langchain_openai import ChatOpenAI
from langchain.agents import create_openai_tools_agent, AgentExecutor
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder

tools = [strut_grounding_tool]
llm = ChatOpenAI(model="gpt-4o-mini", temperature=0)
prompt = ChatPromptTemplate.from_messages([
    ("system", "You are an autonomous research agent using Strut AI web grounding."),
    MessagesPlaceholder(variable_name="chat_history", optional=True),
    ("user", "{input}"),
    MessagesPlaceholder(variable_name="agent_scratchpad"),
])

agent = create_openai_tools_agent(llm, tools, prompt)
executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

# Run Query
response = executor.invoke({"input": "What are the latest developments in quantum qubits in 2026?"})
print(response["output"])

import { openai } from '@ai-sdk/openai';
import { streamText, tool } from 'ai';
import { z } from 'zod';

export async function POST(req: Request) {
  const { messages } = await req.json();

  const result = streamText({
    model: openai('gpt-4o-mini'),
    messages,
    tools: {
      strutGround: tool({
        description: 'Ground live web queries with sub-second verified factual citations',
        parameters: z.object({
          query: z.string().describe('The web search query to ground factually')
        }),
        execute: async ({ query }) => {
          const res = await fetch('$STRUT_BASE_URL/v1/ground', {
            method: 'POST',
            headers: { 'Content-Type': 'application/json' },
            body: JSON.stringify({ query, max_sources: 4 })
          });
          return await res.json();
        }
      })
    }
  });

  return result.toDataStreamResponse();
}

from sdk.python.strut import StrutClient

client = StrutClient(base_url="https://your-deployment.example.com", api_key=os.environ["STRUT_API_KEY"])

# 1. Factual Grounding
grounded = client.ground("What is topological quantum computing?")
print("Answer:", grounded["grounded_answer"])

# 2. Fast AI Search
search_res = client.search("Latest AI breakthroughs 2026", search_depth="fast")
for item in search_res["results"]:
    print(f"• {item['title']} ({item['domain']}): {item['snippet'][:100]}...")

# 3. Batch Markdown Extraction
extracted = client.extract(["https://en.wikipedia.org/wiki/Quantum_computing"])
print("Markdown preview:", extracted["results"][0]["markdown"][:300])