stomwerk content, answered.

stomwerk content is a creative production suite for marketing teams. You teach it the brand once - it ingests your guidelines into structured brand facts - then plan campaigns, generate assets and ship channel-ready packs, with every piece checked against the brand it came from. It runs at engine.stomwerk.com as part of stomwerk scale.

Marketing and brand teams who need to produce more content without losing the brand - the people carrying campaigns, social, and multi-market or multi-language output, and the reviewers who keep it on brand.

A generic model knows nothing about your brand and forgets what you tell it. stomwerk content ingests your brand guidelines into structured brand facts, retrieves them for every generation, checks every claim against them, and routes everything through your review queue. The unit of work is a governed campaign, not a prompt.

Most tools bolt a brand voice onto a prompt. stomwerk content makes the brand the system: generation is grounded in brand facts you approved, claims are checked against them, review is built in, and your brand knowledge, library and assets are yours in EU storage - a capability you own, not a black box you rent.

You upload your brand guidelines and the engine proposes structured brand facts and entities - voice, claims, terminology, visual rules - which you inspect, edit and approve. Generation retrieves from that approved knowledge, and your review decisions keep refining it.

Brand facts are the structured, approved statements the engine must obey - voice, claims you can make, terminology, visual rules. Entities are the people, products and concepts the brand talks about. Together they are working brand knowledge the engine retrieves and checks against, not a PDF nobody opens.

When the engine produces content, the claims it makes are checked against your approved brand facts rather than left to chance. A claim that is not supported is flagged, so you are not shipping something the brand cannot stand behind.

Assets arrive ranked and pre-checked against brand facts, so reviewers approve, revise or reject rather than rewrite. Those approval signals teach the engine your taste over time, so the queue gets closer to on-brand with use. A person approves everything that ships.

A broad set of asset types - text posts, images, carousels, documents and short motion - planned as campaigns and packaged per channel. It also produces grounded research dossiers to base campaigns on, kept with the work they informed.

Once you approve an asset, it fans out into other formats, channels and languages. The next channel or market costs minutes rather than a new project, and everything derived carries the same approved brand and claims.

Yes. Approved assets can be taken into other languages by transcreation - adapted for the market, not just translated word for word - so the message lands rather than reads like a machine translation.

Channel packs are assets that arrive sized, formatted and packaged per channel, ready to publish, so the work of reformatting for each destination is done by the engine rather than by hand.

Yes. Image generation uses Stability models on Amazon Bedrock. Because Bedrock offers no EU-resident image model, image generation runs in a US region while text generation runs on Amazon Nova in the EU (Stockholm); in both cases prompts are not stored or used for training, and generated images are stored back in EU storage. The full detail is on the subprocessor register.

Assets export in the common formats teams need - images, documents, and video for motion - packaged as channel-ready packs. Your library and its assets are yours to export at any time.

Yes. Every workspace has an AI off switch, and the manual path remains: the asset library, review flows, archive and channel packs all work with human-made assets too.

No. Your brand knowledge and content are not used to train AI models. Generation runs through providers under agreements that prohibit training on your prompts and outputs, and prompts are processed transiently.

In the EU. The application and database run in Frankfurt and assets are stored in EU-jurisdiction object storage. Stomwerk AB is a Swedish company operating under Swedish law and the GDPR; the DPA and subprocessor register are on the legal shelf. The one exception is image generation, which runs in a US region because Bedrock has no EU image model - the resulting images are stored back in the EU.

AI assistance is disclosed, human review is built into every flow, and AI can be turned off - in line with the EU AI Act's transparency expectations. Content is also structured so search and answer engines can cite it accurately.

stomwerk content is the self-serve product on the stomwerk platform: create an account at engine.stomwerk.com, set up your brand, and start producing. Six tiers from EUR 19 to EUR 399 per month (Solo and Team, each in Core, Pro and Ultra), published at stomwerk.com/pricing. Subscriptions renew monthly (or yearly, if chosen) until cancelled - cancel anytime from your account settings.

Saved voice profiles for named colleagues, available on Pro and Ultra tiers. From an approved piece of content, stomwerk content drafts supporting posts in each persona's voice; the persona owner reviews, edits and publishes from their own drafts inbox, and nothing is posted without their say-so.

stomwerk content has six tiers: Solo Core EUR 19, Solo Pro EUR 39 and Solo Ultra EUR 89 for one seat, and Team Core EUR 99, Team Pro EUR 199 and Team Ultra EUR 399 for teams of 5 to 20. Each tier includes a monthly AI allowance and EU storage; extra users (Team tiers), an AI boost and extra storage are add-ons. Current tiers are published at stomwerk.com/pricing.