Writing & MediaFree
Legal Review Prep Before Publication
Send the draft and its evidence, and get every actionable assertion ranked by exposure with its source and reply status attached.
A media lawyer reading a draft in page order asks about whatever comes first, not whatever is riskiest. Confirming the source behind assertion four takes an email; confirming it for assertion nineteen takes another, sent after the first reply arrives, often a day later. The Reporters Committee for Freedom of the Press's own pre-publication review guide defines the exposure precisely: a false statement of fact about an identifiable person, published to a third party, that damages their reputation. Nothing in a page-order draft tells counsel which sentences are actually that.
On a worked draft for The Bellwood Courier, an invented paper investigating a demolition contractor, fourteen assertions name an identifiable person or company in a way that could be actionable if wrong. One is backed only by a single anonymous source and accuses two people, a business owner and a city inspector, of bribery: the single highest-exposure sentence in the piece. In the draft's own paragraph order it does not appear until paragraph twenty-one of thirty-four, after twenty paragraphs of well-documented material that make it read as already established.
This tool separates the fourteen from everything else in the draft. Then it sorts them by what drives exposure: the harm category, a crime, a professional failing, a lesser reputational claim, or a labelled opinion, crossed with how thin the evidence is. It also flags who in that ranked set was never offered a chance to respond, the gap the Reporters Committee's own vetting guidance calls out directly. Run it before the fact check that grades every other assertion for truth, so counsel reads the whole risk surface once, not in four separate passes.
A lawyer's first question is never about paragraph one
The map sorts on two axes, not one. The category of harm: a crime, a professional failing, a lesser reputational claim, or an opinion clearly labelled as one. And the evidence behind it: a document, a named on-the-record source, an anonymous background source, or nothing at all. On the worked draft, fourteen assertions split six professional, four other reputational, three opinion and one crime; five rest on a document, four on a named source, three on background, two on nothing at all. Crossed, one sentence, a crime on background alone, outranks all thirteen others.
The same sentence can carry two different legal tests. On the worked draft, one line accuses a private business owner and a public building inspector of the same bribery. The owner, a private figure, only needs to show the paper was negligent to win a libel claim. The inspector, a public official, has to show the paper knew the claim was false or strongly suspected it, the actual malice standard the Reporters Committee's own explainer walks through. A map that does not name which subject is which cannot tell counsel which standard actually applies here.
Two more fields ride along with each row: whether a document could carry the fair report privilege, and whether the named subject was ever asked to respond. On the worked draft, the owner was contacted twice and declined; the inspector, named in the single riskiest line, was never asked at all. That gap is the kind of finding a multi-day, page-order review can miss until deadline. None of this is legal advice, and it does not decide what runs. It prepares the read counsel still has to make on their own.
How it works
Send the file
Attach the draft and the evidence behind it, then say who was and wasn't asked to respond.
Separate the risk
River pulls out only the assertions that name someone and could harm them if wrong.
Cross and rank
Each one gets a harm category and an evidence tier, then the set is ranked by the two combined.
Read the map
A Doc orders every entry by exposure; a Sheet holds the same rows for counsel to sort and mark up.
What you get
- Splits every assertion in the draft into actionable and not, before anyone reads a word of it
- Names the harm category for each one: a crime, a professional failing, another reputational claim, or opinion
- Grades the evidence behind it as a document, a named source, a background source, or nothing at all
- Flags every subject who was never offered a chance to respond, ranked alongside the claim it belongs to
- Names each subject's public or private status, since that changes which fault standard actually applies to the line
- Builds the ranked Doc and the sortable Sheet from the same rows, so neither drifts from the other
Common questions
Is this the same as a fact check?
No. Fact check grades every assertion in the piece against its source for truth. This tool only pulls the subset that could be legally actionable if false, then orders that subset by exposure for counsel. Run fact check for accuracy and this for risk; a draft usually needs both before it runs.
Does it decide what's defamatory?
No, and it should not be read as legal advice. It applies a consistent, visible rule for ranking risk, so counsel's own judgment call is the first thing they make on each line rather than the fifth question in an email thread. The Reporters Committee's own guide is explicit that no vetting process replaces an attorney.
Why does the subject's public or private status matter?
Because it sets which legal standard applies to that specific line. A public official or public figure has to prove the newsroom knew a claim was false or strongly suspected it; a private figure only has to prove negligence, a far lower bar. On the worked draft, the same sentence names one of each, so it carries two different standards at once.
What counts as a potentially actionable assertion?
One that names or clearly identifies a real person or company and, if false, would harm their reputation: an accusation of a crime, incompetence, dishonesty or a similar reputational claim. A neutral fact, like a date or a location, is not actionable even if it is wrong. The Reporters Committee's guide uses the same test: false, about an identifiable person, and damaging.
What if the riskiest claim is backed by a document?
Then it usually carries less exposure, not more, because a fair and accurate account of an official record or proceeding can carry the fair report privilege in many states. The map marks which rows are privilege candidates so counsel checks that first, rather than reading the whole draft to find them. Confirming the privilege actually applies is still theirs to decide.
Where does this sit in the rest of the workflow?
After reporting and before the piece runs, alongside a date the editorial calendar should already hold open for it. It reads the same evidence file the investigation timeline builds and the same response record the no-surprises letter produces. The newsroom's own standards policy governs the sourcing and conflict calls behind the draft. If a complaint arrives anyway after publication, the correction pack is next, not this.
Legal Review Prep Before Publication
Fill in the form and your workspace opens with the work already underway.