Content generation endpoints

Generate visibility-optimized content, score SEO, and improve pages via the REST API.

Base path: /v1/ai/generate

These endpoints support the content side of your AI visibility workflow — creating and improving pages that help AI models mention and cite your brand. For visibility reporting, see AI visibility endpoints.

POST /text

Generate content from a prompt. Use this to create pages targeting prompts from your visibility report.

Body:

Field	Type	Required	Description
text	string	Yes	Prompt / user message
model	string	Yes	Model ID (OpenAI, Anthropic, or Gemini family)
systemPrompt	string	No	Custom system instructions
brandVoiceId	string	No	Apply a saved brand voice
brandVoice	boolean	No	Enable brand voice context
file	string	No	Attachment reference for context
useWorker	boolean	No	Queue async job instead of synchronous response
jobType	string	No	Worker job name when useWorker is true (default: ai_tool_content)
options	object	No	Extra options passed to the worker

Sync response (200):

{
  "message": "Generated AI successfully.",
  "data": "Generated text content..."
}

Async response (202) when useWorker: true:

{
  "message": "Content generation queued",
  "jobId": "12345"
}

Example:

curl -X POST https://api.krastie.ai/v1/ai/generate/text \
  -H "Content-Type: application/json" \
  -H "api-key: krastie_YOUR_KEY" \
  -d '{
    "text": "Write an FAQ page answering: What are the best tools for AI search visibility?",
    "model": "gpt-4o",
    "brandVoiceId": "BRAND_VOICE_ID"
  }'

POST /chat

Conversational AI with workspace visibility context. Enable useWorkspaceContext to include your latest visibility scores, traffic data, and documents.

Body:

Field	Type	Required	Description
text	string	Yes	User message
model	string	Yes	Model ID
sessionId	string	No	Existing chat session (created if omitted)
systemPrompt	string	No	Custom system prompt
useWorkspaceContext	boolean	No	Include visibility and workspace data (default: true)
brandVoiceId	string	No	Brand voice context
webpageContext	string	No	Raw page text for context
webSearch	boolean	No	Enable web search (OpenAI models only)

Response (200):

{
  "message": "Generated AI successfully.",
  "sessionId": "abc123",
  "data": {
    "content": "Assistant reply...",
    "title": "User message"
  },
  "contextSources": []
}

POST /calculate-seo-score

Score content for SEO before publishing visibility-optimized pages.

Body:

Field	Type	Required	Description
content	string	Yes	Body text to analyze
title	string	No	Page title
keywords	string[]	No	Target keywords from your visibility report
autoExtractKeywords	boolean	No	Extract keywords when none provided (default: true)

POST /improve-content-simple

Single-pass AI content improvement with before/after SEO scores. Use on existing pages that underperform in visibility reports.

Body:

Field	Type	Required	Description
content	string	Yes	Content to improve
improvePrompt	string	Yes	Improvement instructions
metricType	string	No	Metric label (default: general)

POST /improve-content-with-validation

Iterative content improvement with validation. Requires a paid plan.

POST /process-blog-post-with-images

Process a blog post and generate associated images.

POST /image

Generate an image.

Body:

Field	Type	Required	Description
text	string	Yes	Image prompt
model	string	Yes	dall-e-3 or stability-ultra
size	string	Yes	DALL·E size (e.g. 1024x1024) or Stability aspect ratio

Response (200): Provider-specific image payload in data.

Brand voice (generate routes)

Create brand voices that keep visibility-optimized content on-brand.

Method	Endpoint	Description
POST	/brand-voice/website	Generate brand voice from a website URL
POST	/brand-voice/text	Generate brand voice from sample text
POST	/brand-voice/file	Generate brand voice from an uploaded file

Manage saved voices via Brand voice library.

POST /feedback

Submit feedback on AI output quality.

GET /job/:jobId

Poll async job status (content, SEO, or AI queues).

Response (200):

{
  "jobId": "12345",
  "name": "ai_tool_content",
  "state": "completed",
  "progress": 100,
  "result": {},
  "failedReason": null,
  "attemptsMade": 1,
  "processedOn": 1717776000000,
  "finishedOn": 1717776005000,
  "queue": "aiQueue"
}

Job states follow BullMQ conventions: waiting, active, completed, failed, etc.