ngram review for UK small businesses

ngram is an AI video creation platform for business teams. You give it source material — a written prompt, a PDF, a deck, a public web page, a screenshot or a screen recording — and it plans the video scene by scene before generating anything. The workflow is deliberately agentic: describe the video, review and approve the scene plan, watch the build, compare the preview against the storyboard, then make point-and-change edits in chat that are pinned to individual scenes alongside a traditional timeline editor.

The generation layer produces the parts that would otherwise need several tools or a freelancer: script, storyboard, AI visuals, AI voiceover and background music, motion graphics, captions and subtitles. Editing is credit-based for AI operations at one credit per second, with unlimited manual edits, chat-based scene regeneration, and smart zoom and auto-cut tools for cleaning up screen recordings. Brand kits, custom voices and custom avatars keep output consistent, and the platform advertises eye-contact correction for presenter-style footage.

Distribution covers the formats UK businesses actually use. Exports run from 720p to 4K depending on plan in widescreen, vertical and square aspect ratios; translation and voice dubbing span more than 25 language hubs with lip sync; and hosted video with an embeddable player is in limited rollout, with beta social publishing to LinkedIn, X and YouTube. Engagement features — polls, quizzes and calls to action inside hosted playback — sit on the top Ultimate plan.

Identity deserves care when reading this review, because the name is genuinely ambiguous. This is Ngram, a California corporation contracting under California law with arbitration in San Francisco, running a PartnerStack partner programme, led by Anish Muppalaneni and Devadutta Ghat. It is not the Google Books Ngram Viewer, not the statistical n-gram language-model concept, and not the life-sciences product that previously occupied the same domain before the company pivoted. It is also not any of the similarly named small agencies and tools that appear in UK searches.

UK Business Stack classifies ngram as the video production layer of a content stack: it sits where an agency brief, an editing tool and a captioning workflow would otherwise overlap, and it produces assets that then need human review for accuracy, brand fit and accessibility before publication.

ngram occupies the space between a written brief and a finished video file. Where a team would normally commission a freelancer, brief an agency or spend hours in an editing suite, ngram plans the scenes, generates the assets and returns an editable project.

The integration story is lighter than the feature list suggests. Automation runs through Zapier, n8n and viaSocket with Make.com announced but not yet available, an MCP server lets AI agents create videos, and a Chrome extension and Google Workspace add-on capture source material. There are no native connectors for Slack, Salesforce or common project tools, so embedding video into those systems relies on embed code.

The UK question is about the vendor rather than the capability. The contracting entity is a California corporation under California law, pricing is in US dollars only, no UK entity or UK support team was found, and the privacy policy refers data transfers to standard contractual clauses without offering UK or EU residency — a standard set of considerations for a thirteen-person company.

The flow below reflects ngram's documented product behaviour and published feature pages. UK Business Stack has not run a live ngram trial for this review, so treat these steps as documented behaviour rather than tested behaviour.

Before subscribing to any AI video platform it is worth naming the honest alternatives. All of them produce video; they differ in input style, creative control, price and how much human editing remains.

• Who you are contracting with. The contracting entity is Ngram, a California corporation, with governing law in California and arbitration in San Francisco; the team of about thirteen people is distributed across the United States and India with no UK entity found.
• Currency and tax. All published prices are in US dollars — free, $29, $59 and $299 monthly tiers with annual discounts, and Enterprise from $30,000 per year — with no GBP price list, so sterling cost moves with exchange rates; confirm VAT treatment at checkout.
• Where the data sits. The privacy policy says data may be transferred to and processed in other countries with standard contractual clauses where required, and does not publish a UK or EU residency option; the sub-processor list is available only on request.
• UK GDPR position. The terms reference UK GDPR and make a data processing addendum available on request, which is the right instrument to ask for, but the practical transfer and residency arrangements still need to be confirmed in writing.
• Support in UK hours. Support runs through a help centre, community forum, documentation and live chat with a 24 business hour response target on the contact form; no UK office, phone line or UK-hours commitment was found.
• Vendor maturity and continuity. The company is small, its founding year is reported inconsistently across its own and third-party sources, and it pivoted from an earlier product on the same domain — all reasonable points of diligence for a business-critical subscription.
• Accessibility and consent obligations. Captions support accessibility requirements, but custom avatars and voices rely on consent rules you must satisfy yourself, and synthetic presenters carry reputational considerations in UK client-facing content.
• Integration fit for a UK stack. Zapier, n8n and viaSocket cover automation and a Google Workspace add-on covers documents, but native connectors for Salesforce, Slack and common UK SaaS tools are absent.

ngram uses credit-based monthly pricing in US dollars: a Free plan with 600 one-time credits and a watermark, Basic at $29 per month ($23.20 billed annually) with 3,000 credits and 720p export, Premium at $59 ($47.20 billed annually) with 10,000 credits and 1080p, Ultimate at $299 ($239.20 billed annually) with 50,000 credits and 4K, and Enterprise custom from $30,000 per year.

Credits are the unit of real cost: AI generation starts at 10 credits per second and AI editing operations cost one credit per second, so a minute of generated video consumes meaningful credits. Top-ups are published at 1,000 credits for between $10 and $40 depending on plan, which means effective per-minute costs vary with tier and usage.

All prices are in US dollars with annual billing discounts of about twenty per cent, and there is no sterling price list. Budget holders should confirm current prices, top-up rates and the VAT treatment that applies at checkout before committing.

• No UK entity or sterling pricing. The contracting entity is a California corporation under California law, all prices are in US dollars, and no UK entity was found — routine friction for UK procurement, finance and VAT handling.
• No published UK or EU data residency. Transfers rely on standard contractual clauses and the sub-processor list is available only on request, so residency and processing locations must be confirmed in writing for UK data protection obligations.
• A very small company. With roughly thirteen employees, support capacity, continuity and roadmap stability carry more risk than with established vendors, particularly for teams making ngram a critical workflow.
• Credit costs are hard to forecast. Generation at 10 credits per second, AI editing at one credit per second and translation all draw from one monthly pool, so actual cost per video depends on length and how much rework happens.
• Distribution features are unfinished. Hosted video and embed playback are described as a limited rollout and social publishing as beta, so core distribution may still change after you build a process around it.
• Thin native integration coverage. There are no native connectors for Slack, Salesforce or common project tools, and Make.com is announced but not live, so automation depends on Zapier, n8n or manual embedding.
• Output still needs human review. Generated claims, visuals, pronunciation and captions are not verified by the platform, so publishing without editorial review risks factual and brand errors reaching your audience.