Introduction
What Crowdee is, how it works, and where to start.
Crowdee is a hybrid content verification platform that combines AI analysis with human-in-the-loop crowd review. It processes text, images, audio, video, and multimodal content — and produces auditable verdicts, not just confidence scores.
Crowdee GmbH is headquartered in Berlin. All infrastructure runs in Germany, under EU data protection law.
How It Works
Upload your content
Upload a file (image, audio, video, text, or PDF) to a project, or provide a URL or text snippet directly as context.
Choose a verification pipeline
Select one of 14 pre-built pipelines based on your media type and the depth of analysis required. Pipelines range from fast AI-only checks (~1 min) to full AI + expert crowd reviews (~150 min).
AI runs the first analysis pass
Every pipeline starts with one or more AI stages that extract signals, assess technical integrity, and identify anomalies.
Crowd reviewers add human judgment (Tier 2 and 3 pipelines)
For deeper verification, vetted crowd workers and domain experts review the AI findings and submit structured responses. Consensus thresholds gate each crowd stage.
Receive an auditable verdict
The pipeline produces a final verdict (authentic, manipulated, synthetic, inconclusive, or unverified) with a confidence score and a stage-by-stage audit trail.
Key Concepts
Organizations & Projects
How billing, modules, and workspaces are structured.
Files & Enrichment
How files are uploaded, stored, and enriched before verification.
Verification Pipelines
The 14 pipelines, their tiers, and how to choose the right one.
API Authentication
API keys, JWT Bearer tokens, and OIDC login.
Explore by Use Case
Verify an image
Start with fast metadata analysis or go deep with the full image authentication pipeline.
Fact-check an article
Parse claims, assess source credibility, and route to fact-checker crowd workers.
Detect audio deepfakes
Combine AI spectral analysis with native-speaker crowd review.
Check URL credibility
Assess domain reputation and content context in ~2 minutes — no file upload needed.
Trace content to its source
Live web research plus crowd fact-checker confirmation to find where content actually came from.
How is this guide?