Skip to main content
The AI content endpoints turn ideas, text, or media URLs into draft social copy and images. They use the same generators as the Postsiva app, WhatsApp agent, MCP, and GPT Actions.
None of these endpoints publish to social networks. Chain with Posting (POST /unified/post/*) or the MCP publish tool when you are ready to go live.

Authentication

All routes require a workspace API key or JWT session.
Every endpoint also requires workspace_id as a query parameter (UUID). With an API key, the keyโ€™s workspace is authoritative โ€” workspace_id must match if you send both. AI generation consumes AI credits on your plan (generation_long = 2 credits, image_generate = 5 credits per platform run).

Idea to content

Generate platform-specific post copy from a topic or idea. Mirrors MCP tool idea_to_content.

POST /unified/content/generate

Query parameters Request body * idea can be empty if every requested platform has an entry in idea_by_platform. Supported platforms: linkedin, instagram, facebook, bluesky, threads, tiktok, youtube, pinterest, or all.
Response example
object
Per-platform keys. Each value is either { "content": "..." } with full post text, or { "error": "..." } on failure for that platform.
Per-platform ideas example

Content to image

Generate images from post copy. Mirrors MCP tool content_to_image.

POST /unified/image/generate

Query parameters Request body * content can be omitted if every platform has content_by_platform. Supported platforms: all, linkedin, instagram, facebook, threads, tiktok, pinterest, twitter (alias x), bluesky. Platforms run in parallel.
Response example
UUID
Postsiva unified media ID. Pass to POST /unified/post/image as default_image_id when publishing.

Image to content

Generate post copy from a single image URL. Mirrors MCP media_to_content with media_type=image.

POST /unified/image-to-content/generate

Query parameters Request body Supported platforms: linkedin, instagram, threads, tiktok, pinterest, all.
Response example

Video to content

Generate post copy from a video URL. Mirrors MCP media_to_content with media_type=video.

POST /unified/video-to-content/generate

Query parameters Request body Supported platforms: linkedin, instagram, threads, tiktok, pinterest, youtube, facebook, bluesky, all.
Response example

Analyze media (MCP / agent only)

Describe an image or video without generating post copy. Available via:
  • MCP tool analyze_media
  • GPT Actions POST /unified/gpt-actions/analyze_media
  • Workspace agent chat (POST /workspace-agent/website/chat)
There is no dedicated REST route for vision-only analysis. Use analyze_media when you need a neutral description before drafting, and media_to_content (or the image/video-to-content endpoints above) when you want publish-ready captions.

End-to-end pipeline

Example flow
  1. POST /unified/content/generate โ€” draft LinkedIn + Instagram copy
  2. POST /unified/image/generate โ€” create images from the LinkedIn caption
  3. POST /unified/post/image โ€” publish with default_image_id from step 2

MCP equivalents


Errors

Generated copy may exceed platform character limits. Trim or rephrase before publishing โ€” the posting endpoints enforce limits per network.

Next

Posting

Publish generated copy and images

Agent chat

Natural-language agent with the same tools

MCP tools

Tool catalog for AI agents