Image-Generation

Image-Generation — ɳSelf documentation.

Feature: Image Generation

ɳClaw supports multi-provider image generation. You can generate images from text prompts through the chat interface or the API, and browse results in the image gallery.

Route: /gallery in claw-web API: POST /claw/image/generate

Supported Providers

Provider	Env Var	Notes
DALL-E (default)	`OPENAI_API_KEY`	OpenAI’s image model
Stable Diffusion	`SD_API_URL`	Self-hosted or API
Midjourney	`MJ_API_KEY`	Via proxy API

Set CLAW_IMAGE_PROVIDER in your .env to choose the default provider. Users can override per-request.

How It Works

Send a prompt via chat or POST /claw/image/generate
The request enters the job queue (np_claw.image_jobs table)
The selected provider processes the request
Results appear in the gallery and in the chat thread

API

Generate Image

POST /claw/image/generate
Content-Type: application/json
Authorization: Bearer {token}

{
  "prompt": "A sunset over mountains in watercolor style",
  "provider": "dall-e",
  "size": "1024x1024"
}

List Jobs

GET /claw/image/jobs

Returns all generation jobs with status (queued, processing, complete, failed).

Get Job

GET /claw/image/jobs/{id}

Returns job details including the generated image URL when complete.

Gallery

The /gallery route in claw-web shows all generated images as cards. Each card displays the prompt, provider, timestamp, and a zoomable preview.

Configuration

Env Var	Default	Description
`CLAW_IMAGE_PROVIDER`	`dall-e`	Default provider
`OPENAI_API_KEY`	—	Required for DALL-E
`SD_API_URL`	—	Stable Diffusion endpoint
`MJ_API_KEY`	—	Midjourney proxy key

Feature-ɳClaw — ɳClaw overview
Feature-Agent-Dashboard — Agent metrics