News and publishing
A news archive any agent can query
The archive exposed so an editor can ask it questions.
Archivequeried by question, not by keyword
AI think tank · newsrooms and media · Mumbai, working worldwide
DigiDesk brings aggregation, writing, publishing and media onto one desk — built by the research team that wrote the models underneath it.
Why this one works
Language, vision and audio, done in-house.
The research ends up somewhere an editor works.
On-premise, your cloud, or ours.
There is no generic configuration.
The platform
DigiDesk is where our newsroom research ends up. Six modules, taken together or one at a time.
The platform
An editorial platform, not another point tool. Aggregation, assisted writing, publishing, media handling and reader-facing widgets, brought onto one desk so a story does not change systems four times between tip-off and publication.
It is the accumulation of our research in news media, which is why the parts fit together rather than merely integrate.
See what DigiDesk doesPulls in from every source you watch and surfaces what an editor should see, so the output is a shortlist rather than another feed to read.
A drafting tool that works alongside a writer on research, structure and revision. The judgement stays where it belongs.
Approved draft through to live, with the automated steps sitting inside the workflow your desk already runs.
Everything you have shot or photographed, labelled and searchable by what is actually in it rather than by filename.
One story, several languages — translation, transcription and dubbing without a separate production run for each.
Recommendations, audio versions of articles and analytics that drop into the site you already have.
Built in-house
Six more products, each out of a real deployment, each running on infrastructure you control.
Ask your own news archive a question and get an answer, not a result list.
Makes a video library searchable, cheaper to deliver, and safe to open up.
Gets more out of the cameras you already own, without replacing any of them.
An email assistant that runs on your server, inside the client you already use.
Search your archive by describing it, or by who is in it. Documents, photographs and video.
Makes the cameras you already have understand what they are seeing — covered faces included.
Research areas
Six areas, one team. The products above are what came out of them.
Selected work
News and publishing
The archive exposed so an editor can ask it questions.
Archivequeried by question, not by keyword
Travel and hospitality
Shore-side booking built from scratch, on-board IT and satellite links overseen — as CIO of a national cruise operator.
30+engineers led across the programme
Education technology
A tutor students can talk to. Nothing left the network.
100%on-premise, no external calls
Insights
On a defence site the question is not whether cloud AI is good enough. Outbound traffic is not permitted, so the system has to be complete inside the perimeter.
Cameras that only help after the fact are not doing much. The system should watch every feed and raise the alert before the incident — the person then decides.
A law firm, a hospital or a defence unit cannot hand a vendor standing access to its mail. OpenMailBot runs on your own server, in the client you already use.
Keyword search fails because nobody named the file for the question you are asking now. DeepFile indexes what is in the material — text, images and video.
Most AI vendors need your data to travel to them. For many organisations that is not a trade they can make — so we build systems with nowhere to send it.
A call about the problem, not a pitch about the platform. If AI is the wrong answer we will say so.
Scoping only: what you are trying to solve, and how to reach you. It will not answer general questions, write code, or quote prices.
Hello — I help scope projects for Manotr. Tell me what you are trying to solve and I will ask a few questions, then pass a summary to the team. What is the problem?
AI assistant, limited to scoping a project with Manotr. It can make mistakes and does not speak for the company. Your conversation is emailed to us so we can follow up. Prefer email? ankit@manotr.com