NextBlock Cortex AI is the AI layer built for the way NextBlock pages are actually composed. Instead of treating your site like a blank document, it understands blocks, sections, editor constraints, and the model choices your team wants to use.
Model routing
Pick the right model
Route generation through OpenRouter or your configured provider strategy so each task can balance speed, quality, and cost.
BYOK control
Use your own keys
Keep provider credentials under your control while still giving editors a clean, guided AI workflow in the CMS.
Typed output
Generate valid blocks
Schema-aware generation helps AI output land as usable page structure instead of loose copy that needs a rebuild.
Why Cortex AI Belongs Inside the Editor
Generic chat tools can draft copy, but they do not know the difference between a hero, a card grid, a product description, and a localized article. Cortex AI lives closer to the editing surface, so generation can respect the same block contracts that render on the public site.
That makes AI useful for practical production work: drafting a landing section, tightening an excerpt, expanding a product story, translating a post, or reshaping a rough idea into a block layout that already fits the system.
Editorial workflow
Faster first drafts
Start with a structured prompt and get a section, article outline, or product narrative that already matches the site voice.
Cleaner revisions
Ask for shorter, clearer, more technical, more polished, or locale-ready output without leaving the content screen.
Model Routing and Cost Control
Cortex AI is designed around provider choice. Teams can route requests through the configured AI gateway, choose task-appropriate models, and keep bring-your-own-key setups separate from the editor experience. Editors get a simple control surface while developers keep the operational knobs.
- Use fast, inexpensive models for drafts, variations, and rewrites
- Reserve stronger models for long-form strategy, technical copy, or difficult transformations
- Keep provider keys and routing defaults in the server-side configuration layer
- Make cost and quality decisions without changing the public rendering system
Block-Aware Generation
The important shift is that Cortex AI is not just writing text. It is meant to produce content that can map back to NextBlock surfaces: section copy, article bodies, product descriptions, headings, calls to action, and localized variants. That reduces the cleanup step that usually happens after copying AI text from a separate tool.
Because the output is shaped around existing block contracts, the generated content feels native in the editor and predictable on the frontend.
Safer Team Workflows
AI works best when it is helpful without becoming invisible infrastructure. Cortex AI keeps humans in the loop: editors review the output, developers control the available providers, and the CMS keeps the generated content inside the same revision and publishing flow as everything else.
A Practical Launch Flow
- Draft the article, landing section, or product story from a focused prompt.
- Refine the result against the brand voice and target audience.
- Generate a localized version or shorter excerpt for cards and metadata.
- Review the content in the NextBlock editor, publish, and keep iterating through normal revisions.
Cortex AI turns the CMS into a more capable production surface: not a replacement for editorial judgment, but a faster way to shape good ideas into structured pages that are ready to ship.