Developer Reference

Acuityio Developer Hub & API Reference

Integrate audit-grade claim verification into your LLM pipelines, autonomous agents, and content workflows. Choose between sub-second extraction, 5-second gates, or 90-second deep adversarial investigations.

Quickstart Guide Remote MCP Server Error Handbook

POST /extract — Atomic Claim Extraction

1,000+ extractions/day · Latency: ~500ms

Takes unstructured raw text, strips rhetorical fluff and emotional filler, resolves pronouns (e.g., “he” → “Albert Einstein”), and outputs clean, testable factual claims.

cURL Request

curl -X POST https://api.acuityio.online/v1/extract \
  -H "Authorization: Bearer acuity_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Albert Einstein won the 1921 Nobel Prize in Physics for his theory of general relativity."
  }'

POST /assess — Fast 3-Model Gate

Synchronous · Latency: ~5–10s

Runs a fast 3-model consensus gate to check for consensus without waiting for deep debate. Perfect for interactive chat applications and pre-send runtime filters.

cURL Request

curl -X POST https://api.acuityio.online/v1/assess \
  -H "Authorization: Bearer acuity_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "claim": "Albert Einstein won the 1921 Nobel Prize in Physics for his theory of general relativity."
  }'

POST /verify — Deep 8-Model Adversarial Pipeline

Asynchronous (Webhooks) · Latency: ~60–90s

Executes the full 5-stage pipeline: live research across primary sources, 2-round Pro vs. Con debate, 3-expert panel review, and synthesized audit-grade conclusion with 20+ verbatim citations.

cURL Request

curl -X POST https://api.acuityio.online/v1/verify \
  -H "Authorization: Bearer acuity_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "claim": "Albert Einstein won the 1921 Nobel Prize in Physics for his theory of general relativity.",
    "webhook_url": "https://yourapp.com/webhooks/acuity"
  }'

Official Client Libraries & Connectors

Acuityio SDKs

from acuityio import Acuityio

client = Acuityio(api_key="acuity_live_your_api_key")

# 1. Extract atomic claims from unstructured text
claims = client.extract(
    "Albert Einstein won the 1921 Nobel Prize in Physics for his theory of general relativity."
)

# 2. Synchronous fast assessment (~5-10s)
assessment = client.assess(claim=claims[0])
print(f"Fast Verdict: {assessment.verdict} ({assessment.score}/10)")

# 3. Deep 8-model adversarial verification (~60-90s)
job = client.verify(claim=claims[0], webhook_url="https://api.yourapp.com/webhooks/acuity")
print(f"Queued Job ID: {job.id}")

# 4. Grounded follow-up question
answer = client.ask(
    verification_id=job.id,
    question="Did the Nobel committee mention relativity in their citation?"
)
print(answer.text)

Model Context Protocol (MCP)

{
  "mcpServers": {
    "acuityio-verifier": {
      "command": "npx",
      "args": ["-y", "@acuityio/mcp-server@latest"],
      "env": {
        "ACUITY_API_KEY": "acuity_live_your_key_here"
      }
    }
  }
}