Skip to main content

Link

To make it easier to interact with your SD instance wherever it is, we've created a linking service that communicates between the SD instance and one to many clients from a single user.

Ideally, this could be used to expose your SD instance to any front-end that then opts to use the Civitai Link, making it possible to generate right from any page that integrates it.

Special Features​

  • Accessible from any site that integrates the link service
  • Can be used from any device, allow you to managing resources and generating images from your beefy desktop while browsing on your mobile device
  • No need to setup special firewall permissions, cors permissions, or API access, just install the extension and establish your link.

Initial Use Cases​

  • Manage the resources on your SD instance
  • Generate on sites using your own SD instance from any device

Integrating Civitai Link

We've provided a brief outline of how it works below and would be happy to meet with you to discuss integration. The best way to reach us is in the Development channel on our Discord server.

Cool! We're excited to hear what you have in mind. The best way to reach us is in the Development channel on our Discord server.

How the Link works

First Connection​

Client->Server: POST LinkInstance
Note over Server: Generates LinkInstance
Server->Client: Returns LinkKey
Client->Server: Join room with LinkKey
Client-->Extension: User copies over LinkKey
Extension->Server: Join room\nwith LinkKey
Note over Server: Upgrades LinkKey
Note over Server: Moves to upgrade room
Server->Extension: Gives new LinkKey
Note over Extension: Stores new LinkKey
Client->Extension: Makes requests in room
Note over Extension: Perform actions
Extension->Client: Responds with status in room
Note over Server: On Disconnect notify room

Following Connections​

Client​

Client->Server: GET LinkInstances
Note over Server: Finds LinkInstances for User
Server->Client: Returns LinkInstances
Note over Client: Selects LinkInstance
Client->Server: Join room with LinkKey

Extension​

Note over Extension: Load stored LinkKey
Extension->Server: Join room with LinkKey

Note: If the LinkKey no longer exists, the extension is disconnected and the stored LinkKey is removed.

Link Commands

Once the link is established, it can be used for a variety of purposes. Here are a few of the commands we'd like to support initially.

🅰️ = Alpha feature 🅱️ = Beta feature

Base Command​

{
  id: 'uuid', // The identifier for the request
  version: 'semantic version', // Version of the request
  createdAt: Date,
  updateAt?: Date
}

activities:list 🅰️​

Request

{}

Response

[{
  ... Response of last 10 requests
}]

resources:add 🅰️​

Request

{
  id: 'uuid',
  resources: [{
    type: 'lora' | 'checkpoint' ...;
    hash: 'sha256',
    name: 'name of file',
    previewImage: 'url to preview image',
    url: 'url to download if missing'
  }]
}

Response

{
  id: 'uuid',
  resources: [{
    //... props from request
    status: 'pending' | 'processing' | 'success' | 'error' | 'canceled',
    progress: 0-100.00
  }]
  status: 'pending' | 'processing' | 'success' | 'error',
}

resources:add:cancel 🅰️​

Request

{
  id: 'uuid', // id of initial add request
  resources: [{
    type: 'lora' | 'checkpoint' ...;
    hash: 'sha256'
  }]
}

Response

{
  id: 'uuid',
  resources: [{
    //... props from request
    status: 'canceled'
  }]
  status: 'success',
}

resources:remove 🅰️​

Request

{
  id: 'uuid',
  resources: [{
    type: 'lora' | 'checkpoint' ...;
    hash: 'sha256',
  }]
}

Response

{
  id: 'uuid',
  resources: [{
    //... props from request
    status: 'pending' | 'processing' | 'success' | 'error',
  }]
  status: 'pending' | 'processing' | 'success' | 'error',
}

resources:list 🅰️​

Request

{
  id: 'uuid',
  types: [
    'checkpoint',
    'lora',
    //...
  ]
}

Response

{
  id: 'uuid',
  resources: [{
    type: 'lora' | 'checkpoint' ...;
    hash: 'sha256',
    name: 'name of file',
    path: 'optional path to file on fs'
  }]
}

image:txt2img 🅱️​

Note: It's likely that the connection will be upgraded to WebRTC before doing image generation requests to reduce latency and remove our service as a middleman Request

{
  id: 'uuid',
  quantity: 1, // How many to generate
  batchSize: 1,
  model: hash,
  vae: hash,
  additionalNetworks: [{
    type: 'hypernet' | 'lora',
    hash: 'hash',
    name: 'name',
    strength: 0.000-1.000
  }],
  params: {
    prompt,
    negativePrompt,
    sampler,
    seed,
    steps,
    width,
    height,
    cfgScale
  }
}

Response

{
  id: 'uuid',
  images: ['base64']
}

image:img2img 🅱️​

Note: It's likely that the connection will be upgraded to WebRTC before doing image generation requests to reduce latency and remove our service as a middleman Request

{
  //... content of txt2img
  params: {
    //... content of txt2img.params
    initImage: 'url or base64',
    maskImage: 'url or base64',
    denoiseStrength: 0.00-1.00,
    maskedOnly: false
  }
}

Response

{
  id: 'uuid',
  images: ['base64']
}