TNTracenoraSign in

Monitor. Analyse. Act.

Know where the coverage stands.

Tracenora collects news and social posts on the topics you track, scores each one against your organisation’s position, and puts what matters in front of an analyst, in the languages your audience actually speaks.

Sign inSee how it works

Review queue

Topic · Urban air quality

4 pending
  • -2

    Clip claims the city switched off its air monitors before an inspection

    TikTok · English · 41k views in 3 h · no source cited

    Critical
  • -1

    Bus fare rise blamed on new emissions rules

    X · English · High-reach account

    Notable
  • 0

    Wakazi walalamikia moshi wa viwanda karibu na shule

    YouTube · Kiswahili · Residents report factory smoke near a school

    Routine
  • +1

    First month of street-level PM2.5 readings published

    News site · English · Cites the monitoring network

    Routine
Illustrative example. Each item is scored against the topic’s reference position.

The deviation scale

Every item answers one question: how far is it from your position?

For each topic you write down where your organisation stands. Every relevant article, post and video is scored against that position, not against a generic idea of positive or negative.

  1. −2

    Strongly opposed

    Attacks your position or pushes an opposing narrative.

  2. −1

    Divergent

    Frames the topic in a way that contradicts it.

  3. 0

    Neutral

    Factual or descriptive. No clear stance.

  4. +1

    Aligned

    Broadly consistent with your position.

  5. +2

    Strongly aligned

    Explicitly reinforces it.

An angry post can agree with you. A calm, careful report can take your position apart.

That’s why tone and stance are scored separately. Sentiment tells you how something feels. Deviation tells you whether it helps or hurts your position, and it decides what reaches an analyst first.

How it works

From feed to decision.

A pipeline that runs around the clock, so analysts spend their time on judgement, not on reading everything.

  1. 01

    Collect

    Scheduled collectors read news feeds and websites, X, YouTube, TikTok, Instagram and Facebook. Free sources come first; paid ones run only when you switch them on.

    RSS · web pages · X · YouTube · TikTok · Instagram · Facebook

  2. 02

    Filter

    Every item is de-duplicated, its language detected, and matched against your topics’ keywords and hashtags. Weak matches are kept for spot checks, never silently dropped.

    Language detection · keyword and hashtag matching

  3. 03

    Score

    Claude reads each relevant item against the topic’s reference position and returns a deviation score, a triage lane, sentiment, named entities, a short context note and a draft reply in the item’s own language.

    Deviation −2 to +2 · Routine / Notable / Critical

  4. 04

    Review

    Analysts work a queue sorted critical-first: confirm, override, escalate, discard or keep monitoring. Every AI suggestion and every human decision is logged.

    Full audit trail

  5. 05

    Act

    Confirmed critical items trigger email and Telegram alerts. Daily and weekly digests, analytics, strategy reports and a live display wall keep everyone else informed.

    Alerts · digests · analytics · display wall · AI assistant

Who it’s for

Teams that answer for a position.

Nothing about a country, a language or a point of view is built in. Each workspace describes its own organisation, and that shapes every score, alert and digest.

Government & public bodies
See how policy and public-health messaging lands, and catch misinformation before it sets.
Companies & brands
Spot reputation risks around products, recalls and leadership while they are still small.
NGOs & non-profits
Follow how your cause is framed in local and international media, on every platform.
Newsrooms
Track narratives and coordinated messaging across feeds and social video for your desk.
Election observers & researchers
Follow claims about the electoral process and flag incitement in the languages people actually use.

Configured per workspace:

  • Organisation & type
  • Country & region
  • Languages & translation
  • Timezone & date format
  • Entity watchlist
  • Context for the AI

The AI suggests. The analyst decides.

Every score arrives with a context note, every override is recorded, and no alert goes out until an analyst confirms the item.

Sign in

Invited? Use the link in your invitation email to set your password.