EachLabs Provider Management

Manage the lifecycle of EachLabs AI providers: discover new models via API and implement them as IMG.LY providers.

Key Advantages Over Other Providers

EachLabs provides a structured API for model discovery:

1. Programmatic model listing - No need to scrape documentation
2. request_schema - Each model includes JSON Schema for inputs (similar to fal.ai)
3. Consistent API patterns - Unified prediction/polling interface across all models

Key Files

• Provider Tracking: specs/providers/eachlabs/providers.md
• Provider Specifications: specs/providers/ (schemas, architecture)
• Provider Implementations:
  • Image: packages/plugin-ai-image-generation-web/src/eachlabs/
  • Video: packages/plugin-ai-video-generation-web/src/eachlabs/
• README Documentation:
  • Image: packages/plugin-ai-image-generation-web/README.md
  • Video: packages/plugin-ai-video-generation-web/README.md

EachLabs API Reference

Base URL

https://api.eachlabs.ai

Authentication

• API Key via header: X-API-Key: <your-api-key>
• Discovery endpoints do NOT require authentication:
  • GET /v1/models - List all models
  • GET /v1/model?slug=<slug> - Get model details including request_schema
• Prediction endpoints require API key:
  • POST /v1/prediction - Create prediction
  • GET /v1/prediction/{id} - Get prediction status

Endpoints

List All Models (No Auth Required)

GET /v1/models?limit=500&offset=0&name=<filter>

Response:

{
  "models": [
    {
      "title": "Flux 2 Pro",
      "slug": "flux-2-pro",
      "version": "0.0.1",
      "output_type": "image"
    }
  ]
}

Get Model Details (No Auth Required)

GET /v1/model?slug=<model-slug>

Response includes request_schema:

{
  "title": "Flux 2 Pro",
  "slug": "flux-2-pro",
  "version": "0.0.1",
  "output_type": "image",
  "request_schema": {
    "type": "object",
    "required": ["prompt"],
    "properties": {
      "prompt": {
        "type": "string",
        "default": "",
        "description": "The prompt to generate an image from."
      },
      "image_size": {
        "type": "string",
        "default": "landscape_4_3",
        "enum": ["square_hd", "square", "portrait_4_3", "portrait_16_9", "landscape_4_3", "landscape_16_9"]
      }
    }
  }
}

Create Prediction (Auth Required)

POST /v1/prediction
Content-Type: application/json
X-API-Key: <api-key>

{
  "model": "flux-2-pro",
  "version": "0.0.1",
  "input": { ... }
}

Get Prediction Status (Auth Required)

GET /v1/prediction/{id}
X-API-Key: <api-key>

Response:

{
  "id": "...",
  "status": "success|processing|starting|failed|cancelled",
  "output": { ... },
  "predict_time": 12.5,
  "cost": 0.05
}

Workflow Overview

This skill operates in two phases:

1. Discovery (automatic): Fetch models via API and compare against tracked models
2. Implementation (user-approved): Create provider files for selected models

Phase 1: Discovery

Run the discovery script to get available models:

node .claude/skills/partner-providers-eachlabs/discover-models.mjs

This outputs a minimal JSON array with slug, title, and output_type for each model. Compare against specs/providers/eachlabs/providers.md to identify new models.

For detailed steps, see DISCOVERY_CHECKLIST.md.

Model Categories

Classify models by output_type:

• image - Image generation (text-to-image, image-to-image)
• video - Video generation (text-to-video, image-to-video)
• audio - Audio generation
• text - Text generation
• array - Multi-output (e.g., multiple images)

Determining Capability from Slug/Schema

Pattern	Capability
-text-to-video	text-to-video
-image-to-video	image-to-video
image_url in required	image-to-image or image-to-video
Only prompt required, output=image	text-to-image
Only prompt required, output=video	text-to-video

Discovery Output Format

## Discovery Results

### New Models Found
| Slug | Name | Type | Capability | Recommendation |
|------|------|------|------------|----------------|
| flux-2-pro | Flux 2 Pro | image | t2i | implement |

### Updated Models
| Slug | Change |
|------|--------|
| ... | ... |

### Summary
- X new models found
- Y recommended for implementation
- Z skipped (older versions, niche use cases)

After presenting results, ask: "Which models would you like me to implement?"

Phase 2: Implementation

Only proceed after user approval. Follow IMPLEMENTATION_CHECKLIST.md to:

1. Fetch model details to get request_schema
2. Convert request_schema to OpenAPI format (add x-imgly-* extensions)
3. Create provider TypeScript file and JSON schema
4. Export provider from eachlabs/index.ts
5. Add translations to translations.json
6. Add to example app (examples/ai/src/eachlabsProviders.ts)
7. Update specs/providers/eachlabs/providers.md status to "implemented"
8. Update README documentation
9. Run build checks: pnpm --filter "@imgly/plugin-ai-*" check:all

Schema Conversion

EachLabs request_schema is JSON Schema format. Convert to OpenAPI with IMG.LY extensions:

Input (EachLabs):

{
  "type": "object",
  "properties": {
    "prompt": {
      "type": "string",
      "description": "The prompt"
    },
    "image_size": {
      "type": "string",
      "enum": ["square", "landscape_4_3"]
    }
  }
}

Output (OpenAPI with extensions):

{
  "openapi": "3.0.0",
  "info": { "title": "EachLabs API", "version": "1.0.0" },
  "components": {
    "schemas": {
      "ModelNameInput": {
        "type": "object",
        "properties": {
          "prompt": {
            "title": "Prompt",
            "type": "string",
            "description": "The prompt",
            "x-imgly-builder": { "component": "TextArea" }
          },
          "image_size": {
            "title": "Format",
            "type": "string",
            "enum": ["square", "landscape_4_3"],
            "x-imgly-enum-labels": {
              "square": "Square",
              "landscape_4_3": "Landscape 4:3"
            },
            "x-imgly-builder": { "component": "Select" }
          }
        },
        "x-fal-order-properties": ["prompt", "image_size"],
        "required": ["prompt"]
      }
    }
  },
  "paths": {}
}

UI Component Mapping

JSON Schema Type	x-imgly-builder component
string (long text)	TextArea
string (enum)	Select
string (url/image)	Custom via renderCustomProperty
boolean	Switch
integer/number	Number

Status Values

When updating providers.md:

Status	When to Use
implemented	Provider code exists and works
planned	Will implement (high priority)
skipped	Intentionally not implementing (older version, limited use)
undecided	Needs discussion

References

Skill Files (this folder)

• discover-models.mjs - Discovery script (run to find new models)
• DISCOVERY_CHECKLIST.md - Step-by-step discovery process
• IMPLEMENTATION_CHECKLIST.md - Step-by-step implementation process

Provider Specifications

• specs/providers/README.md - Overview of provider system
• specs/providers/architecture.md - How providers fit into the plugin system
• specs/providers/patterns/ui-guidelines.md - CRITICAL: Which parameters to expose/hide in UI
• specs/providers/patterns/text-to-image.md - T2I implementation pattern
• specs/providers/patterns/image-to-image.md - I2I implementation pattern
• specs/providers/patterns/text-to-video.md - T2V implementation pattern
• specs/providers/patterns/image-to-video.md - I2V implementation pattern

EachLabs-Specific

• specs/providers/eachlabs/providers.md - Model inventory and status