GoFactAI

Integration

How integration works

Fact AI Lab sits as a thin verification layer between your application and your LLM provider. The platform offers a Python SDK, a TypeScript SDK, and a REST API that works with any OpenAI-compatible endpoint.

You do not change your prompts or your model. You pipe outputs through the platform before they reach users. Average integration time for a single workflow is under 2 hours.

What you send

Each verification request includes:

  • The LLM output you want verified.
  • The source documents or context chunks the LLM was given.
  • Optional configuration: workflow ID, retention policy, score threshold, and escalation behavior.

The platform cross-references the output against the source documents to produce a verification score and a list of flagged claims.

What you receive

Each response includes:

  • The verification score (0.0-1.0, where 1.0 means every claim is fully supported).
  • A list of flagged claims with their source citations or lack thereof.
  • An audit log entry ID for the tamper-evident record written to your storage.

If the score falls below your configured threshold, the platform can suppress the output, return a configurable fallback string, or route the interaction to a human review queue.

The audit log entry

Every verified interaction produces a log entry containing:

  • The original prompt (sanitized or full, your choice).
  • The retrieved context chunks with source citations.
  • The model version and temperature.
  • The raw LLM output.
  • The verification score.
  • Any flagged claims.
  • A cryptographic hash linking to the previous entry.

Logs are tamper-evident and can be exported as PDF or JSON for regulatory review.

Access and security

API keys can be rotated on-demand without opening a support ticket. IP allowlisting is available for API access. Webhook events include signature verification. Detailed access logs are available for your own SIEM.

Regions

Data processing region is configurable: US, EU, or CA. Audit logs are written to your S3-compatible storage in your designated region. No cross-region data movement occurs without explicit configuration.

For data handling details, see the Security Model. For questions about setup and workflow configuration, see the FAQ.