# Multimedia Verification (/docs/products/multimedia-verification)



Multimedia Verification is Crowdee's core product: 14 pre-built pipelines that combine AI analysis with optional crowd review to detect manipulated, forged, or AI-generated content across every modality.

## What It Covers [#what-it-covers]

* **Technical integrity** — file metadata, compression artifacts, and format anomalies (Tier 1, AI-only, \~1–3 minutes).
* **Deep manipulation and deepfake detection** — pixel-level forensics, temporal tampering, and audio spectral analysis, escalated to crowd review for consequential decisions (Tier 2, AI + crowd, 40–150 minutes).
* **Synthetic content detection** — a dedicated verdict, distinct from "manipulated," for content that's AI-generated from scratch rather than an edited original. Covered across image, video, audio, and text pipelines.
* **Context and multi-file cases** — URL credibility, identity claims, and cross-modal corroboration, which don't fit the single-file model (Tier 3).

Every run produces a structured scorecard and a stage-by-stage audit trail, whether or not a crowd stage was involved.

## Cost [#cost]

500–8,900 credits per run, depending on tier and modality — a Tier 1 technical check costs far less than a Tier 2 AI + crowd deep review. Each pipeline's exact `costCredits` is listed in the [pipeline catalog](/docs/pipelines).

## Where to Go Next [#where-to-go-next]

* [Verification Pipelines](/docs/pipelines) — the full catalog of 14 pipelines, organized by tier, with a lookup table by content type.
* [Verdicts](/docs/concepts/verdicts) — what each verdict means and how confidence scores are computed.
* [Synthetic Data Detection](/docs/concepts/synthetic-data-detection) — how the `synthetic` verdict differs from `manipulated`, and which pipelines support it.
* [Pipeline Stages](/docs/concepts/pipeline-stages) — how AI, crowd, and web-research stages compose into a single pipeline run.