Submitting Content for Certification

Content Certification accepts documents in multiple formats through both dashboard and API. This guide covers submission methods, supported formats, and size limits.

Submission Methods

Dashboard Submission

1. Navigate to Certification → Submit Content
2. Choose submission method:
   • Paste Text: Copy-paste content directly into text editor
   • Upload File: Upload PDF, DOCX, HTML, or markdown files
   • URL: Provide link to publicly accessible content (auto-fetches text)
3. Add metadata:
   • Title (optional): Display name for this certificate
   • Tags (optional): Categorize by type (marketing, product, legal, etc.)
   • Batch ID (optional): Group with other submissions for tracking
4. Click Submit for Verification

API Submission

Single Document:

from truthvouch import TruthVouch

client = TruthVouch(api_key="your-api-key")

response = client.certification.submit(
    content="Your AI-generated content here...",
    content_type="text",
    title="Product Guide v2.1",
    tags=["product", "marketing"],
    batch_id="batch-launch-2024-q1"
)

print(response.certificate_id)

Multiple Documents (Batch):

documents = [
    {
        "content": "Content 1",
        "title": "Guide A",
        "tags": ["product"]
    },
    {
        "content": "Content 2",
        "title": "Guide B",
        "tags": ["marketing"]
    }
]

batch = client.certification.batch_submit(
    documents=documents,
    batch_id="batch-123"
)

print(f"Batch ID: {batch.id}")
print(f"Documents submitted: {batch.document_count}")

Supported Formats

Text Formats

• Plain Text (.txt, .md)
• HTML (.html, .htm) — removes HTML tags, preserves structure
• Markdown (.md) — processes markdown syntax

Document Formats

• PDF (.pdf) — extracts text with optional OCR
• Microsoft Word (.docx) — extracts content, preserves headers
• Google Docs (via URL) — auto-fetches and extracts

Constraints by Format

Format	Max Size	Processing	Notes
Plain Text	10 MB	<10s	Fastest option
HTML	5 MB	10-20s	CSS/JS stripped
PDF	5 MB	20-60s	OCR applied if needed
DOCX	5 MB	15-30s	Headers/footers skipped
URL	N/A	20-40s	Must be publicly accessible

Content Guidelines

What Works Well

• Product Specifications: Detailed features, pricing, compatibility
• Marketing Copy: Claims about company, products, capabilities
• Process Documentation: Standard operating procedures, policies
• Legal Agreements: Terms of service, privacy policies
• Technical Guides: API documentation, configuration guides

What Doesn’t Work Well

• Opinion-Based Content: Personal perspectives, recommendations (not fact-checkable)
• Purely Narrative Text: Stories without verifiable claims
• Image-Only Content: PDFs containing only images without text
• Real-Time Data: Stock prices, weather, live system status (changes too frequently)

Preparation Tips

1. Structure Your Content: Use headers, lists, and short paragraphs
2. Avoid Hallucinations: Manually review before submission (certification catches but doesn’t prevent)
3. Link to Truth: Reference specific truth nuggets in your knowledge base
4. Date Your Content: Include publication date for time-sensitive claims

Batch Submission

Batch submission is ideal for certifying multiple documents at once:

Dashboard Batch

1. Navigate to Certification → Batch Submit
2. Click Add Documents → Select multiple files
3. Configure batch settings:
   • Auto-Revoke On Change: Automatically revoke if facts change
   • Expiry Duration: When certificates expire (30/60/90 days, 1 year)
   • Notification Settings: Who gets alerts on completion
4. Click Submit Batch

API Batch

import json

# Prepare batch CSV
batch_data = [
    {"id": 1, "content": "Document content 1", "type": "product"},
    {"id": 2, "content": "Document content 2", "type": "marketing"},
]

# Submit batch
batch = client.certification.batch_submit_csv(
    data=batch_data,
    batch_id="q1-2024-content",
    auto_revoke_on_drift=True,
    expiry_days=90
)

# Poll status
while batch.status != "completed":
    import time
    time.sleep(10)
    batch = client.certification.get_batch(batch.id)

# Retrieve results
results = client.certification.batch_results(batch.id)
for result in results:
    print(f"Doc {result.id}: Score {result.trust_score}, Status: {result.status}")

Monitoring Progress

Automatic Status Updates

Dashboard:

• Navigate to Certification → Recent Submissions
• Automatically view submission status and estimated completion time
• Click to preview automatically extracted claims

API:

# Get submission status
submission = client.certification.get_submission(submission_id)
print(f"Status: {submission.status}")
print(f"Progress: {submission.processed_claims}/{submission.total_claims}")

# Get batch status
batch = client.certification.get_batch(batch_id)
print(f"Batch Status: {batch.status}")
print(f"Completed: {batch.completed_count}/{batch.total_count}")

Handling Errors

If submission fails:

1. Unsupported Format: Ensure file is in supported format list
2. Size Exceeded: Split large documents into smaller sections
3. Encoding Issues: Ensure text is UTF-8 encoded
4. No Verifiable Claims: Content may not contain fact-checkable assertions

View detailed error messages in dashboard or via API:

submission = client.certification.get_submission(submission_id)
if submission.status == "failed":
    print(f"Error: {submission.error_message}")

Next Steps

• Verification Process: Learn how claims are extracted and verified
• Trust Score: Understand what influences certificate scores
• Monitoring: Set up alerts for certificate changes