NNxtedition AI Agents

In September 2026, Nxtedition showed AI agents taking a Reuters story from the wire to the studio gallery without leaving its own system. The agents drafted the story, pulled video from the media store, applied graphics templates and built the running script, and the show then ran from the gallery. A project built on the same platform for BBC India, a nine-language newsroom with transcription and translation on the broadcaster’s own hardware, made the shortlist for an IBC2026 Innovation Award.

Nxtedition AI Agents suit TV news broadcasters that want AI working across scripting, video, graphics and rundowns, and that care where their data goes: the models can run on the broadcaster’s own GPUs with no per-query cost. The agents work best when the newsroom runs on Nxtedition’s platform, and pricing is by quote. We rate it 7.6/10 on our scale. We have not used Nxtedition ourselves; this review is based on the company’s announcements and trade press coverage.

Two Trade Show Seasons of Agent Announcements

Nxtedition is a broadcast software company rather than an AI startup, and its agent story has built up over two trade show seasons. These are the facts we could confirm on the source pages.

Two things matter for a buyer. First, the agents are part of the production system, not a bolt-on. They can reach the same storage, rundowns and gallery controls that staff use, which is why Nxtedition can show a single story moving from wire copy to air without exports. Second, the cost model is unusual for AI: the company has said since 2025 that AI processing carries no extra fees, and the 2026 on-premise option removes per-query charges entirely.

Public evidence is the next thing to build. Outside the BBC India project, the public material is mostly trade show demonstrations and company statements. Ask for reference broadcasters running agents in daily production, not only at a stand.

A Story-Centric Newsroom and Production Platform

Nxtedition is a story-centric newsroom and production platform: journalists plan stories, write scripts, edit video and build rundowns in one place, and directors run the show from the gallery. The agents work inside that platform and call the same back-end services.

The platform uses a microservices architecture, so operators and studios can work from different locations for remote production. It handles video up to 4K and HDR, offers fast-turnaround editing, integrates with DaVinci Resolve and exposes an open API that already connects tools such as MapCreator for dynamic maps and the Melodie production music library. Agents inherit that reach, which is why they can move a story through editing, graphics and rundown without exporting files to another product.

From Newswire to Gallery

• Propose: the AI suggests angles, new stories and breaking news to journalists, using sources such as Reuters.
• Develop: agents draft the story and pull relevant video from the media store.
• Package: they apply graphics templates and assemble a running script with prompter text, video clips and graphics.
• Review: journalists and directors check and edit every script and video before publication.
• Air: the show is recorded in the studio and run from the gallery inside the same system.

Background Work at Ingest

Live transcription subtitles and tags content as it is ingested and flags sections for automatic clipping, running on the broadcaster’s own GPUs. Since 2025, agents have also drafted titles and descriptions for social posts from news stories or video transcripts, suggested new angles from current scripts, fact-checked incoming material and adapted stories for different audiences.

Models and Deployment

At NAB 2025 Nxtedition said its agents combined several open-source models, naming OpenAI Whisper and Meta’s Llama 3, which also power semantic natural language search. By IBC2026 the company described the setup as model-agnostic: it ships GPU hardware with on-premise language models and can call external models where that suits a broadcaster better. Deployment can be on premises or hybrid.

Control and Oversight

Every agent step appears in the story planning view and the rundown, so staff can see, change or discard what the AI did. Broadcasters can also drive the agents from an external AI assistant through MCP, which opens the system to outside tools without custom integration code.

Editorial Judgment and Gallery Control Stay Human

Editorial judgment stays with people. Journalists approve scripts and video, directors run the gallery, and staff decide which AI-proposed stories and angles to pursue. Someone also has to manage the GPU hardware and model choices in an on-premise setup.

Quoted Broadcaster Contracts Without List Prices

Nxtedition sells its platform to broadcasters under quoted contracts and does not publish list prices. What it has said about cost concerns the AI layer.

Comparing Against Per-Minute AI Services

With no list prices, the useful comparison is against per-minute or per-query AI services a newsroom might otherwise pay for. The volumes below are our own illustrative assumptions.

• A 24-hour news channel ingesting 12 hours of feeds a day: that is 4,380 hours of video a year transcribed, subtitled and tagged. On Nxtedition’s on-premise setup, none of those hours carries a per-query charge; multiply by your current per-minute transcription rate to see what that is worth.
• A newsroom publishing 40 stories a day to three social platforms: agents drafting a title and description for each post would produce 120 drafts a day, or 43,800 a year, with no per-call AI fee.
• A multi-language operation like BBC India: running nine languages of transcription and translation on local hardware shifts cost from usage fees to the upfront GPU purchase and its upkeep.

Costs to plan beyond the license: GPU hardware for on-premise models, migration of the newsroom and media store onto the platform, staff training, and integration with existing graphics, playout and publishing systems.

Nxtedition AI Agents Compared With Amagi Newspulse, Moments Lab and Descript Underlord

AI agents for news video fall into three groups: full production systems with agents built in, specialist tools that sit beside an existing newsroom, and creator editing tools. Nxtedition is the first kind.

A broadcaster replacing its newsroom system can look at Nxtedition for the whole job. One keeping its current system will find it easier to add Amagi Newspulse for social output or Moments Lab for archive search. Descript is the low-cost choice for teams making recorded video rather than live news. See more in our AI agents for media and video guides.

TV Newsrooms Replacing Their Production System

• TV news broadcasters replacing a newsroom system: the agents come with a platform that covers scripting, editing, rundowns and gallery control.
• Multi-language newsrooms: the BBC India project shows nine languages of transcription and translation on one platform.
• Broadcasters with strict data rules: on-premise models keep content in the facility, with no data leaving the building.
• High-volume channels worried about AI usage bills: the no-per-query model makes heavy transcription and drafting predictable.
• Teams that want outside AI assistants to act on the newsroom: MCP support lets an external assistant drive the agents.

Broadcasters Staying on Another Newsroom System

• Broadcasters committed to another newsroom system: the agents work inside Nxtedition, so the value depends on moving production onto it.
• Podcasters and social video creators: a self-serve editor such as Descript fits the job and budget better.
• Teams without IT staff for on-premise hardware: GPU servers in the building need someone to run them.
• Publishers working mainly in text: the system is built around video, rundowns and studio production.

Planning for an All-in-One Newsroom System

• Platform fit: agents, storage, rundowns and gallery control all live in one system, which is the point of the product and means a move to it is a long-term decision.
• Demo versus daily use: the full newswire-to-gallery chain was shown at IBC2026; ask how much of it reference customers run every day.
• Hardware ownership: no per-query cost means the broadcaster pays for and maintains the GPUs instead.
• Editorial accountability: AI-suggested angles and fact-checks still need verification; Nxtedition keeps journalists and directors in the approval loop for that reason.
• Pricing by quote: there is no public price list, so compare total cost including hardware and migration.

Reuters as News Partner for the IBC Demo

• On September 7, 2026, IAMT reported that Nxtedition would demonstrate agentic AI from newswire to gallery at IBC2026, with Reuters as news partner and the BBC India Collective Newsroom named an Innovation Awards finalist.
• In April 2025, Nxtedition showed AI Agents and AI Commands at NAB, running open-source models including Whisper and Llama 3 with no extra processing fees.
• At the same show, Nxtedition let visitors control its automated XR studio gallery with a headset, alongside the new agents.

Among the Most Complete Options for Single-Platform Newsrooms

For a broadcast newsroom ready to run production on one platform, Nxtedition AI Agents are among the most complete options shown so far: they act across the whole chain, from wire copy to the gallery, and every step stays visible and reversible in the rundown. The on-premise model with no per-query cost is a real difference for channels that transcribe and clip around the clock. It scores 7.6/10 on our scale, with quote-based pricing, a substantial setup and public evidence beyond the BBC India project as the areas to build.

Treat the IBC2026 workflow as the direction of travel and test the parts you need now, such as ingest transcription, clipping and social drafts. Ask for reference broadcasters, a full cost picture including GPUs, and a pilot on one bulletin before committing.

Models, Cost and Whether Content Leaves the Building

How Much Do Nxtedition AI Agents Cost?

Nxtedition does not publish prices, so broadcasters request a quote for the platform. The company has said AI processing carries no additional fees, and its on-premise option ships GPU hardware with local models and no per-query cost.

Which AI Models Does Nxtedition Use?

At NAB 2025 the company named open-source models including OpenAI Whisper and Meta’s Llama 3. By IBC2026 it described the setup as model-agnostic, running local language models on the broadcaster’s GPUs and calling external models where that suits a broadcaster better.

Does Content Leave Our Building?

Not on the on-premise setup. Nxtedition says its local models run on the broadcaster’s own hardware with no data leaving the facility, and it also offers hybrid deployment where external models are preferred.

Can the Agents Put a Show on Air Without Staff?

No. The agents draft stories, pull video, add graphics and build the script, but journalists and directors review and edit every script and video before publication, and the gallery still runs the show.

How Does Nxtedition Compare With Amagi Newspulse?

Amagi Newspulse turns existing newscasts into clips, vertical video and bulletins, while Nxtedition is the production system itself, with agents helping to build the show. A broadcaster keeping its newsroom system will find Newspulse easier to add; one replacing it can consider Nxtedition.