Skip to main content
Content Studio · Fact Checking

Every claim, checked.
Every verdict, sourced.

Paste any text and Helix extracts the factual claims, cross-checks each one against authoritative sources, and returns a verdict you can act on — with the evidence to back it up.

fact_check.completed94% confidence
City council unanimously approved the $2B downtown rezoning plan.
Verified4 sources

Council approved the rezoning 7-0 on Tuesday, per the official meeting record.

fayobserver.com

Inputs

Bring your own sources,
or let Helix find them.

Hand it the receipts you already trust, or just paste the text and Helix discovers and cross-references authoritative sources for you.
Hand it your sources

Paste the text plus the URLs you want it checked against — primary documents, official filings, or your own archive.

nasa.gov/mission/apollo11PubMed studiesCourt filingsCity council minutes
Or just paste the text

Skip the sources and Helix discovers and cross-references authoritative ones for you — useful when you do not have time to gather them yourself.

A reporter’s draftA viral social postA politician’s quoteAn AI-generated summary
How it works

Extract, verify, deliver the verdict.

A reviewable pipeline, not a black box. Every claim is extracted, every verdict is backed by an evidence trail you can audit.
  1. 01Extract claimsAI identifies factual, verifiable statements and filters out opinion and speculation.
  2. 02Gather sourcesUses the URLs you provide, or auto-discovers authoritative ones if you skip them.
  3. 03VerifyCross-references each claim and classifies evidence as supporting, contradicting, or neutral.
  4. 04Score & deliverReturns a verdict per claim, an overall determination, and confidence scores you can act on.
What you get

A verdict you can actually defend.

Not a vague pass/fail. Every result comes with the underlying claims, the evidence, and a confidence score per claim and overall.
Verdict per claim

Verified, disputed, mixed, or unverified — with the evidence balance the verdict was built on.

Evidence with attribution

The exact passages that backed the verdict, linked to the source page they came from.

Confidence scores

Per-claim and overall, so editors know what to trust at a glance and what needs a human.

The payload

Submit, retrieve, subscribe.

A single endpoint to submit, one to retrieve claims with evidence, and a webhook to fire on completion.
curl -X POST https://api.feeds.onhelix.ai/fact-check \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"text": "The Moon landing occurred in 1969. Neil Armstrong was the first person to walk on the Moon.",
"sourceUrls": [
"https://www.nasa.gov/mission_pages/apollo/missions/apollo11.html"
]
}'
Where it ships

Wire it into your editorial flow.

Drops into the review step you already have. Use the API directly, or fire on webhook completion into your CMS.
In the field

Built for the people who cannot afford to be wrong.

One review step, every editorial workflow.
For editorial desks
Catch the bad claim in the draft

Run every draft through fact-check before publish. Disputed claims surface for human review; verified ones move through clean.

For radio & podcasts
Vet what goes on air

Run quotes, statistics, and script passages through verification in seconds — with the contradicting evidence right there.

For research desks
Audit at scale

Fact-check hundreds of pieces in one pass and surface only the low-confidence ones for a human pass.

Don't ship what
you can't defend.

Get in touch

Forty-five minutes with a Helix engineer and an editor who has shipped what you are trying to ship. No sales calls.