FactScopeFactScope

AI fact check reports in minutes

FactScope orchestrates multi-agent workflows and adaptive model routing to break down claims, retrieve evidence, grade sources, and ship cited verdicts. Submit any text for a fact check and receive both structured JSON and a ready-to-render report.

Pay per claim · 1 credit for the standard model

placeholder hero

Five agents, one reliable fact check

Input moderation, claim extraction, retrieval, evidence grading, and report generation each produce structured artifacts so every FactScope task is reproducible.

  • Input safety review
    Use omni-moderation-latest to block disallowed content while reusing the same request ID throughout the pipeline.
  • Claim extraction & query building
    Collaborative models break the text into verifiable statements and craft 3–6 diverse search queries.
  • Trusted-source retrieval
    Multi-engine recall, then whitelist filtering for gov/edu/journals and tiered media sources.
  • Evidence grading & preliminary verdict
    Source tiers, relevance scores, and temporal checks feed structured verdicts with confidence.
  • Structured report delivery
    Verdict Composer aggregates evidence, qualifiers, and citations into a JSON contract easy to render or archive.
Trust & Safety

Built for high-trust fact check workflows

From retrieval to reporting, FactScope enforces source weighting, temporal validation, and structured outputs so every fact check stays explainable.

Prioritize gov, edu, peer-reviewed journals, and tier-one media while tagging evidence levels.

Source whitelist
Temporal validation
Structured report
Use Cases

Cover the most common editorial fact checks

Default retrieval tactics and citation templates help you fact check policy, data, and event statements across content formats.

Short-form script reviews

Fact check policy claims, numbers, or events before publishing video scripts.

Brand statements & PR

Fact check claims before external announcements to avoid spreading unverified statements.

Community escalations

Quickly fact check user reports and provide auditable rulings with cited evidence.

Encyclopedia & knowledge-base updates

Fact check whether entries are outdated and highlight the missing evidence types.

Where FactScope helps fact check teams

We focus on text today yet already cover claim extraction, evidence retrieval, verdicting, logging, and billing across the full fact check pipeline.

Claim extraction + automatic queries

Generate search queries with synonyms, counter-claims, and temporal filters straight from user text.

Dynamic model routing

Select across SOTA models by risk level and enable dual verdicts when claims conflict.

Async background execution

Return a task ID instantly and poll status instead of blocking long HTTP sessions.

Quality guardrails

Source tiers, temporal checks, and conflict alerts run together; low-signal cases fall back to "unclear" with auto refunds.

Structured JSON output

Verdicts, confidence, recommended phrasing, and citations follow one schema ready for front-end rendering or database writes.

Cost control

Optimize reasoning spend while holding quality thresholds so per-claim pricing stays predictable.

Performance

Shaped by real fact check workloads

Current release supports concurrent jobs, lifecycle logging, and usage analytics so you always know what each fact check report costs.

Automatic claim discovery

≤5 claims

Adaptive upper bound (soft 3, hard 5) depending on submission length.

Background concurrency

2 jobs

Two claims run in parallel by default to balance latency and spend.

Evidence citations

≤3 refs

Show up to three top-ranked citations per claim with caching and reuse.

Choose the FactScope credit pack that fits

Pick a Basic, Pro, or Max bundle with recurring credits. Every credit unlocks one full Simple fact check workflow run—from moderation through verdict JSON.

Basic

Solo fact check
$13/month

220 credits every month for occasional fact checks. Credits expire when the next month begins.

Includes

  • 220 credits ≈ 220 standard claims per month
  • Credits expire at the start of the following month
  • Simple workflow: moderation → claims → retrieval → verdict JSON
  • Console dashboard plus downloadable audit logs

Upgrade whenever you need more throughput—unused credits do not roll over.

Pro

Most popular
$22/month

650 credits per month for steady editorial pods that clear dozens of claims each day.

Includes

  • 650 credits ≈ 650 claims per month
  • Simple workflow with cited verdict JSON for every claim
  • Console dashboard plus team-visibility usage logs
  • Automatic refunds on blocked or failed runs

Best value for teams running 500–700 claims monthly.

Max

Newsroom scale
$59/month

2,200 credits every month for high-volume fact check desks that need guaranteed throughput.

Includes

  • 2,200 credits ≈ 2,200 claims per month
  • Simple workflow from moderation to verdict JSON
  • Console plus exportable billing & performance logs
  • Ideal for newsroom-scale queues and automation hooks

Work with our team to pre-plan surges or newsroom-wide deployments.

FAQ

Quick answers about FactScope’s capabilities, pricing, and safeguards.

1

Why do I need to sign in before submitting?

Fact checking consumes credits and logs usage, so we must link every task to an account for billing and audit.

2

What if multiple claims are detected?

We show the claim list and projected spend first; you confirm before running and no credits are charged for unapproved claims.

3

Do I pay if evidence is inconclusive?

If the pipeline completes and the verdict is "unclear," the task still counts; failed or blocked runs refund the held credits.

4

Do you support other languages or modalities?

We focus on Chinese text today. Additional languages plus audio/video support will roll out in future releases.

Experience the AI fact check pipeline

Submit any text for a fact check and get a cited, auditable report within minutes.

Contact the team