For a single office-holder, or for a whole team?
Both, and the difference is not size but governance. An office-holder with one staffer needs one person to sign off; a ministerial team has a spokesperson, a chief of staff and a trail that holds up afterwards. That second case is why the approval flow is configurable rather than fixed.
| Who | Who decides | What keeps them up at night |
|---|---|---|
| Office-holder | The office-holder themselves, with one staffer | Is this allowed — because it must not be campaigning — and is what circulates about me accurate |
| Ministerial team, group or party staff | The chief of staff or the spokesperson | Governance, a trail that holds up afterwards, and coverage across several people at once |
How does a quote I never gave come into existence?
Almost never out of malice. An engine reads a report where your name and a position sit in the same paragraph, and turns it into one sentence. That is not a lie but a merge — and exactly why the source underneath is the only thing that helps: without an address there is nothing to correct.
| Driver | What ceeme measures underneath | What stays dark without a connection |
|---|---|---|
| Hallucination risk | invented and outdated facts about you, counted by type | We see what an engine invents about you, and we follow the media coverage around it through the press feed. |
| Reputation risk | misquotes, sliding reviews and where they come from | Reviews and social are read live. Print and paywalled media are not in there. |
| Source authority | the standing of the sources that name you | Which sources name you and how much they weigh, we see. Writing to them is human work. |
| Sentiment / framing | the role you get cast in — the cheap one, the safe one, the fast one | We measure the framing per answer and break it out per source — so you see which source carries the picture. |
| Governance risk | who may publish what, and whether it can be undone | We show who may do what. Enforcing it per channel needs an operator flag and an adapter. |
Do you also track what is said in closed groups?
No. What is closed stays closed: no adapter, no scraper, no route into a private group. What we read is the press, public social and the forums anyone can reach — and of that last group we cover broadly but not exhaustively, as the table says.
| Source | Kind |
|---|---|
| press and news | public — we read it, you connect nothing |
| X | public — we read it, you connect nothing |
| public — we read it, you connect nothing | |
| communities | public — we read it, you connect nothing |
| Wikidata | public — we read it, you connect nothing |
| press feed | connected — only after your approval, read-only unless agreed otherwise |
| social insights | connected — only after your approval, read-only unless agreed otherwise |
Why do you refuse to write campaign content?
Because it is a different trade with different rules, and because mixing the two damages both. Monitoring only works if the outcome is checkable; persuasion only works if it persuades. A system that does both stops being believed in the first.
| Rule | Kind |
|---|---|
| Campaign and persuasion copy does not come out. What does come out is what factually happened, with the source attached. | hard rule — the build refuses output that crosses it |
| Nothing goes out without the person authorised to sign off at your organisation having seen it — the flow sits in the code, not in an understanding. | hard rule — the build refuses output that crosses it |
| Stays outside every route. What you do with a factual frame is up to you and whoever advises you. | forbidden — we do not do this, not even on request |
| We neither track nor write anything about an opponent — that builds material whose only use is campaigning. | forbidden — we do not do this, not even on request |
| No manufactured support, no orchestrated replies, no accounts made to look like something. | forbidden — we do not do this, not even on request |
| No message that differs per audience. What gets published stands the same for everyone. | forbidden — we do not do this, not even on request |
| Every claim carries its source. What has no source is not written — not even when it looks true. | forbidden — we do not do this, not even on request |
| Not inside a ministerial team either, and not as a clause inside a factual correction. | hard rule — the build refuses output that crosses it |
| Every finding, every approval and every change is logged with time and person, and can be read back afterwards. | hard rule — the build refuses output that crosses it |
How do I respond when something false about me is circulating?
You get the finding with its source, and a frame stating what is factually right and what is not. Where it goes is your call: a correction to the newsroom, a clarification on your own channel, or nothing — because sometimes a misquote nobody picks up is better ignored than amplified.
Frequently asked questions
Is this allowed during a pre-election blackout period?
We are not lawyers and this is not legal advice, but the design points that way: nothing is produced that could count as electoral communication, not even on request. What remains is checking whether a statement attributed to you is accurate, and that is fact-checking rather than campaigning. Put it to your own counsel.
Can you track what is being said about my opponent?
No. Opponent content is forbidden in the code, and that includes monitoring — because tracking what is said about someone else builds material whose only use is campaigning. The exception is when someone else's statement causes a misquote of you; then that source is part of your finding.
How complete is the media coverage?
Not complete, and the table above says so. Public press and social are read live; print and paywalled media are not, and of the forums we cover broadly but not exhaustively. A monitoring system that acts as if it sees everything is more dangerous than none — because then you read silence as absence.
Read on
Free, no account and no card.