Frequently Asked Questions

Generative AI optimization is the discipline of shaping prompts, evaluation processes, and inference safeguards so that AI outputs are accurate, useful, and aligned with business goals. It goes beyond basic prompt engineering to include multi-stage verification, brand governance, and feedback loops. J77.ai was built around these principles, with every piece of content passing through research, critique, and verification stages.

Prompt engineering focuses on crafting individual prompts for better outputs from a single interaction. GenAI optimization encompasses the entire content pipeline—including research, multi-model critique, claim verification, brand alignment, and governance workflows. While prompt engineering is one component, optimization addresses the full lifecycle to ensure consistent, accurate, enterprise-grade results.

J77.ai uses a multi-stage verification pipeline: (1) Real-time web research grounds content in current sources, (2) Draft generation with source attribution, (3) Multi-model critique reviews accuracy and logic, (4) Claim extraction identifies factual statements, (5) Each claim is scored as supported, weak, or unsupported based on available evidence. Unsupported claims are flagged, softened, or removed before delivery.

Verified AI content is AI-generated material that has been systematically fact-checked, source-attributed, and validated against real-world evidence. Unlike raw AI outputs, verified content includes source citations, confidence scoring for claims, and traceable evidence chains. This eliminates hallucination liability and provides audit trails for compliance.

Answer Engine Optimization is the practice of structuring content so AI-powered search engines—ChatGPT, Gemini, Perplexity, Claude—can extract, understand, and cite your content in responses. AEO focuses on clear definitions, structured formats (lists, FAQs, tables), semantic richness, and verified accuracy. J77.ai produces AEO-optimized content by default through its structured output and verification pipeline.

Brand-aligned AI content requires systematic governance: (1) Define brand voice, tone, and terminology in reusable Brand Kits, (2) Build a Content Library with approved messaging, CTAs, and glossary terms, (3) Inject this context into every generation automatically, (4) Use multi-model pipelines that enforce consistency. J77.ai's multi-brand architecture ensures every piece of content matches your defined brand personality.

J77.ai uses a multi-model pipeline with multiple specialized AI systems working together. Instead of a single prompt, every piece of content goes through research, drafting, critique, and verification steps. This produces more accurate, original, and well-sourced content than single-prompt tools. The platform implements generative AI optimization principles—verification, governance, and brand alignment—that enterprise teams need.

Each AI handles what it does best: one searches the web for relevant sources, another writes the initial draft with your brand voice, a third critiques and suggests improvements, and then a final merge step incorporates feedback and verifies claims. The result is content that's researched, well-written, and fact-checked.

A Target represents a brand, website, or content destination. It stores your brand voice, tone, SEO preferences, and a library of reusable content. Every build is associated with a Target, ensuring consistency across all your content.

Ten kinds today: blog posts, web pages, news articles, how-to guides, documentation, resource pages, AI images, social-media posts, marketing/outreach emails, and press releases. You can generate single pieces or entire content suites (like a full documentation section) from an outline. Most kinds can also be narrated as audio.

Research Mode searches the web for relevant, recent information before drafting. This ensures your content includes real sources and citations. It's essential for news and trending topics, but can be turned off for internal documentation.

Quality mode runs the full multi-model pipeline on premium models — research, drafting, critique, claim verification, and merge. Use it for anything public-facing or hard to revise. Speed mode is the right pick for drafts and high-volume runs.

A Content Suite generates multiple related pieces of content from an outline. For example, you can paste a documentation structure and J77.ai will create all the pages, with proper internal linking and consistent terminology. You can download the entire suite as a ZIP file.

Set up your brand kit in your Target settings with voice, tone, and guidelines. Add reusable content to your Library (CTAs, product blurbs, glossary terms). The AI uses this context for every build, ensuring consistency.

Yes! Download your content as Markdown or HTML and edit it in any editor. The generated content is a starting point—review it, refine it, and make it yours before publishing.

When Research Mode is enabled, the AI finds relevant sources which are cited in your content. A claim ledger tracks factual statements and their supporting sources. Unsupported claims are softened or removed in the final merge step.

Yes. Your Targets, Library items, and generated content are private to your account. We don't use your content to train models or share it with third parties.

Individual content pieces can be downloaded as Markdown (.md) or HTML. Content Suites are downloaded as a ZIP file containing all files, a navigation structure (nav.json), and an index page.

Yes! You can generate AI images in two ways: enable "Generate Featured Image" when creating a build to get an image for each article, or use the standalone Image Generator to create images on demand. Images are customizable with style, mood, and color settings. You can also overlay your brand logos automatically — upload logos once in Brand Kits → Assets and pick a position. Each image costs 1 token.

Yes. Tick "Include Audio Narration" when creating a build and J77 generates a "Listen" version of each article using Grok TTS (five voices: ara, eve, leo, rex, sal). Audio costs 2 tokens per page and is hosted on J77.ai for streaming or download. Already published an article? Open it from the dashboard and click "Generate Audio" — same flow, no need to re-run the build.

Press releases are a journalist-ready content type with their own structural recipe: dateline, headline + subdeck, lead paragraph answering the 5 Ws, body, power quotes with attribution, boilerplate, "###" terminator, and a media-contact block. Pick "Press Release" as the content type and you get a quote editor that lets you (a) type a quote verbatim, (b) drop in a fragment and have us polish it into final form, or (c) leave it blank for us to generate one attributed to a named speaker. Boilerplate + media contact are stored on your brand kit so you don't retype them every release.

Yes. Connect your WordPress site or Supabase-powered blog in Settings → Blog Connections. Then you can publish or schedule posts directly from J77.ai. Featured images are automatically uploaded to WordPress or included in Supabase posts.

Set up "Relevant Topics" in your Target settings, then click "Suggest Topics" when creating a new build. J77.ai will analyze trending discussions on X (Twitter) and suggest 3 timely blog topics that align with your brand.

Images are hosted on J77.ai's infrastructure as a convenience. For permanent storage, download images to your own hosting. Images may be deleted after 6 months of account inactivity. See the Terms of Use for details.

No, only self-hosted WordPress sites are supported. WordPress.com hosted sites (*.wordpress.com) don't expose the REST API needed for publishing. You need a self-hosted WordPress installation on your own domain.

Yes. J77.ai offers a full REST API that gives you programmatic access to the same multi-model content pipeline available in the web app. You can create targets, start builds, manage your content library, generate images, and more — all via API. Visit the Developers page at j77.ai/developers for details, or explore the interactive Swagger docs at api.j77.ai/api/v1/docs.

Same as the web app. There is no separate API subscription. API usage burns the same token balance as your dashboard builds — buy a token bundle once, spend it from anywhere. The 10 free tokens you get on signup work in the API too.

The API lets you automate your entire content workflow. Common use cases include: batch-creating builds overnight and downloading publish-ready content by morning, managing multiple brand kits for different clients, integrating J77 into your CMS or publishing pipeline, generating images programmatically, and using idempotency keys to safely retry requests. The API uses the same verification pipeline as the web app — every piece of content is researched, critiqued, and claim-verified before delivery.

Yes. J77 exposes an MCP (Model Context Protocol) server that AI assistants can call directly as a tool. OAuth 2.1 with PKCE handles authorization; you grant or revoke access per assistant from Settings → Connections. The OAuth dance is live and verified end-to-end against Supabase; the tool surface (create_build, generate_audio, share_deliverable, etc.) is rolling out this quarter. See j77.ai/developers/mcp for setup details and current status.