Skip to main content
Postcard is a digital forensics tool that takes a social media post URL, traces it back to its original source, and produces a Postcard Score (0–100%) measuring how much the content has drifted from the ground truth.
Trace the Truth.
Screenshots strip context. By the time something goes viral, it’s been cropped, captioned, and misattributed. Postcard uses AI-powered corroboration and live site auditing to restore that context and give you a credibility verdict you can trust.

Quick start

Verify your first post in under a minute

How it works

Understand the 4-stage forensic pipeline

API reference

Integrate Postcard into your own tools

Postcard Score

Learn how the credibility score is calculated

What Postcard does

Postcard runs a four-stage forensic pipeline on any social media post URL:
  1. Ingest — Fetches live content and metadata from the platform using platform-specific strategies (oEmbed, native APIs, or Jina Reader fallback).
  2. Audit — Verifies the post is reachable at its claimed origin using Gemini with Google Search, and checks that timestamps are consistent with the reported narrative.
  3. Corroborate — Searches trusted news and fact-checking domains with Google Gemini to find independent sources that support or refute the content.
  4. Score — Combines origin reachability, corroboration confidence, bias analysis, and temporal alignment into a single weighted Postcard Score.

Supported platforms

Postcard supports direct forensic verification for posts on X (Twitter), Reddit, YouTube, Instagram, Bluesky, and Threads — with generic fallback support for any public URL via Jina Reader.

Submitting a URL

How to start a forensic trace

Reading a report

Understand the verdict and subscores

Supported platforms

Platform coverage and limitations

API key setup

Configure your Gemini API key

Build docs developers (and LLMs) love