4 AI agents · Real-time analysis

Before you share it,
know if it holds up.

Paste any article, claim, or argument. ThinkTrace checks the facts, spots the flaws, and tells you exactly how strong it is — in plain English.

01
Paste anything
Text, a PDF, or a link to an article. No formatting needed — just paste and go.
02
We check it
ThinkTrace reads every claim, looks them up against real sources, and checks the logic.
03
See what's wrong
You get a score, a list of factual errors, and plain-English explanations of any logical flaws.
04
Share or act
Share the report with a link, or use the findings to push back with confidence.
Agents:ParserMapperDetectorVerifier

How it works

What ThinkTrace checks

Parser runs first to extract claims. Then Mapper, Detector and Verifier fire in parallel. Their findings are combined into a single structured report.

01
Parser
Reads your content
Reads your content and picks out every individual claim — the main point and all the supporting statements.
02
Mapper
Builds the argument graph
Draws a map of how the argument is built — which statements back each other up and where the logic breaks down.
03
Detector
Hunts logical fallacies
Spots tricks and weak reasoning — like attacking the person instead of the point, or jumping to conclusions without evidence. Each one is explained simply.
04
Verifier
Fact-checks live
Looks up every factual claim against real sources — Wikipedia, Google, research papers, and news. Tells you what checks out and what doesn't.

Who uses it

Anyone can use it

News and media
Check an article before you share it. Know if the facts hold up.
Academic research
Check your essay or a research paper for gaps in reasoning and unsupported claims.
Legal and policy
Test the strength of a proposal or policy argument before committing to it.
Business decisions
Check a pitch or report for weak claims before making a decision based on it.

Built on

The technology behind ThinkTrace

Want to know how it works under the hood?

AI Orchestration

LangGraph
LangChain
LangSmith

LLM and Agents

Claude API (Sonnet)
4 parallel agents
Epistemic scoring

Async and Queue

Celery workers
Redis broker
Upstash Redis

Data and Memory

PostgreSQL
SQLAlchemy ORM
Neo4j graph
Pinecone vectors

Fact Checking

Serper Search
Wikipedia API
ArXiv + PubMed + News

Backend

FastAPI
JWT + OTP Auth
Multi-tenant orgs

Frontend

Next.js 14
TypeScript
Tailwind CSS

Infrastructure

Docker
Railway cloud
GitHub Actions CI
4
Agents
33
API endpoints
100%
Eval pass rate
5
Fact sources
Railway + GitHub
CI/CD

Built by @Nsk246 · View on GitHub