A truck overturned on the agricultural road; traffic has stoppedA text tip goes straight into analysis: priority, location and a confidence score.
AI for Arabic newsrooms
From thousands of messages to the real story
Nour brings your audience's Facebook and Instagram messages into one inbox, triages them in seconds, and hands every tip to your journalists with a confidence score and a list of what's missing — so they start halfway there, and the call stays theirs.
Built for Arabic & dialect
No automated "confirmed" — journalists decide
No invented details
14,000+real audience messages Nour was built and tested on
88%of inbound isn't news — filtered away from your journalists
<3sto classify a message and estimate its priority
3follow-up questions at most before a tip goes to a journalist
From live use in an Arabic newsroom: 7,607 inbound messages in July–August 2026; median classification time 1.6 seconds.
The problem
The story is in your inbox — buried
Hundreds of messages arrive every day: an urgent tip between dozens of greetings and reactions, a dialect video with no caption, a source who never said where. By the time someone finds it by hand, the scoop is gone.
- Noise drowns the signal
About nine in ten messages aren't news. Someone reads them all to find the tenth.
- Media without context
A dialect video or an uncaptioned photo means nothing until it's opened and watched in full.
- Incomplete tips
Where? When? Is there a photo? The same questions, typed by hand, for every source.
How it works
From an incoming message to a journalist's call
Five steps every message goes through automatically — only what deserves attention reaches your team.
- Arrives
From Facebook Messenger or Instagram; photos and video are archived on the server immediately.
- Triage
News, reply or chatter? Greetings and reactions are dropped without any processing.
- Analysis
Priority, category and location, plus a confidence score and what the tip is missing.
- Follow-up
Ready questions for the source — place, photo, time — three at most.
- The journalist decides
Turn the tip into a story, publish or reject it — and Nour learns from the decision.
Classification
Every message finds its place in seconds
Footage from the fire before the fire brigade arrivedThe audio is transcribed — dialect included — and the frames analyzed, so the story arrives with a summary and transcript.
A photo of a sudden sinkhole on a side streetThe image is analyzed by AI, including any text written inside it.
"Peace be upon you 👍❤️" — a reply to a published storyGreetings and story replies are dropped automatically, without spending any AI.
Capabilities
What Nour handles behind the scenes
Six capabilities working together across text, image and sound — all of it in Arabic.
One sorted inbox
Every conversation in one place, ordered by newest, urgent, or awaiting a reply.
Confidence & what's missing
A score, the number of corroborating reports and the gaps in the tip — never an automated "confirmed".
Video transcripts & image analysis
Speech-to-text that understands dialect, plus analysis of scenes and of text inside images.
Location & coverage map
Governorate and place extracted from the tip, and a map of where reports and sources come from.
Faithful story drafting
Headline, summary and social posts, bound to the source and to your outlet's editorial line.
Nour assistant
Ask in Arabic: it searches every message, surfaces trending topics, and works on each story with you.
Editorial principles
AI that serves journalists — never speaks for them
In a newsroom, being wrong costs more than being late. So Nour is built on three non-negotiable limits.
No automated "confirmed"
Nour estimates confidence, counts corroborating reports and lists the gaps. Judging whether a story is true is the journalist's job alone.
No invented details
Strict rules forbid adding names, numbers, dates or places that aren't in the source. Anything unconfirmed is dropped, not guessed.
Every decision on record
Who published, who rejected, who replied, and when — a full activity log for every member of your team.
Who it's for
One platform for the whole desk
Inbox journalist
Opens a sorted inbox instead of a flood of messages; urgent tips arrive with an instant alert.
Verification editor
Receives each tip with its confidence score, its gaps and its corroborating reports.
Managing editor
Follows team and source performance and the coverage map, and tunes the agents with clear switches.
Channels
Facebook Messenger Supported
Instagram Supported
WhatsApp Coming soon
Read more
The full story
Why we're building Nour, how its technology works, and where it's heading.
Why newsrooms need Nour
Manual triage and what it costs you in scoops.
Read the article
From message to story
A message's journey through the AI pipeline.
Read the article
Nour's AI agents
Their roles, their limits, and the design philosophy.
Read the article
Technology & roadmap
The stack, the channels, and what's next.
Read the articleFAQ
Before you ask
Didn't find your answer? Request a demo and we'll go through integration and privacy in detail.
Does the AI decide a story is "confirmed"?
No — by design. Nour gives a confidence score, the missing details and the number of corroborating reports; the final editorial decision always belongs to the journalist. Our framing is "attribution and confidence", not "automated verification".
Can the AI invent details that aren't there?
Nour is built to prevent it: every draft — headline, summary or post — is bound by faithfulness rules that forbid adding names, numbers, dates or places not present in the source; anything unconfirmed is dropped rather than guessed.
Which platforms are supported?
Facebook Messenger and Instagram via Meta's official APIs, with WhatsApp coming soon. Media is archived on the server so it survives the platforms' expiring links.
Does it understand video and dialect?
Yes. A video's audio is transcribed automatically by a model primed for dialect, and its frames are analyzed, so the journalist gets a classified story with a summary and transcript — without opening the video.
What happens to greetings and story replies?
They're marked "not news" instantly, because they're engagement with the page's content rather than tips — saving journalists' time and AI running costs together.
What happens if the AI goes down?
Nour keeps receiving, storing and triaging every message with fallback rules, alerts your team immediately, and re-processes the backlog once the AI is back. No message is lost.
Demo
See Nour on your newsroom's reality
Send your details and we'll get in touch to set up a live walkthrough.
- A tour of the inbox, triage and the story desk
- Answers on connecting your pages and data privacy
- How Nour is tuned to your outlet's editorial line