Generate Images with Amazon Bedrock

Amazon Bedrock is a fully managed service that offers a choice of high-performing foundation models (FMs) from leading AI companies like AI21 Labs, Anthropic, Cohere, Meta, Mistral AI, Stability AI, and Amazon. Each model is accessible through a common API which implements a broad set of features to help build generative AI applications with security, privacy, and responsible AI in mind.

This guide will walk you through an example using the Amazon Bedrock JavaScript SDK in Supabase Edge Functions to generate images using the Amazon Titan Image Generator G1 model.

Setup#

In your AWS console, navigate to Amazon Bedrock and under "Request model access", select the Amazon Titan Image Generator G1 model.
In your Supabase project, create a .env file in the supabase directory with the following contents:

1
AWS_DEFAULT_REGION="<your_region>"
2
AWS_ACCESS_KEY_ID="<replace_your_own_credentials>"
3
AWS_SECRET_ACCESS_KEY="<replace_your_own_credentials>"
4
AWS_SESSION_TOKEN="<replace_your_own_credentials>"
5
6
# Mocked config files
7
AWS_SHARED_CREDENTIALS_FILE="./aws/credentials"
8
AWS_CONFIG_FILE="./aws/config"

Configure Storage#

[locally] Run supabase start
Open Studio URL: locally | hosted
Navigate to Storage
Click "New bucket"
Create a new public bucket called "images"

Code#

Create a new function in your project:

1
supabase functions new amazon-bedrock

And add the code to the index.ts file:

1
// We need to mock the file system for the AWS SDK to work.
2
import { prepareVirtualFile } from 'https://deno.land/x/mock_file@v1.1.2/mod.ts'
3
4
import { BedrockRuntimeClient, InvokeModelCommand } from 'npm:@aws-sdk/client-bedrock-runtime'
5
import { createClient } from 'npm:@supabase/supabase-js'
6
import { decode } from 'npm:base64-arraybuffer'
7
8
console.log('Hello from Amazon Bedrock!')
9
10
const SUPABASE_PUBLISHABLE_KEYS = JSON.parse(Deno.env.get('SUPABASE_PUBLISHABLE_KEYS')!)
11
12
Deno.serve(async (req) => {
13
  prepareVirtualFile('./aws/config')
14
  prepareVirtualFile('./aws/credentials')
15
16
  const client = new BedrockRuntimeClient({
17
    region: Deno.env.get('AWS_DEFAULT_REGION') ?? 'us-west-2',
18
    credentials: {
19
      accessKeyId: Deno.env.get('AWS_ACCESS_KEY_ID') ?? '',
20
      secretAccessKey: Deno.env.get('AWS_SECRET_ACCESS_KEY') ?? '',
21
      sessionToken: Deno.env.get('AWS_SESSION_TOKEN') ?? '',
22
    },
23
  })
24
25
  const { prompt, seed } = await req.json()
26
  console.log(prompt)
27
  const input = {
28
    contentType: 'application/json',
29
    accept: '*/*',
30
    modelId: 'amazon.titan-image-generator-v1',
31
    body: JSON.stringify({
32
      taskType: 'TEXT_IMAGE',
33
      textToImageParams: { text: prompt },
34
      imageGenerationConfig: {
35
        numberOfImages: 1,
36
        quality: 'standard',
37
        cfgScale: 8.0,
38
        height: 512,
39
        width: 512,
40
        seed: seed ?? 0,
41
      },
42
    }),
43
  }
44
45
  const command = new InvokeModelCommand(input)
46
  const response = await client.send(command)
47
  console.log(response)
48
49
  if (response.$metadata.httpStatusCode === 200) {
50
    const { body, $metadata } = response
51
52
    const textDecoder = new TextDecoder('utf-8')
53
    const jsonString = textDecoder.decode(body.buffer)
54
    const parsedData = JSON.parse(jsonString)
55
    console.log(parsedData)
56
    const image = parsedData.images[0]
57
58
    const supabaseClient = createClient(
59
      // Supabase API URL - env var exported by default.
60
      Deno.env.get('SUPABASE_URL')!,
61
      // Using the default Supabase API PUB KEY.
62
      // If you want to use a different api key, change 'default' to your preferred key name
63
      SUPABASE_PUBLISHABLE_KEYS['default']
64
    )
65
66
    const { data: upload, error: uploadError } = await supabaseClient.storage
67
      .from('images')
68
      .upload(`${$metadata.requestId ?? ''}.png`, decode(image), {
69
        contentType: 'image/png',
70
        cacheControl: '3600',
71
        upsert: false,
72
      })
73
    if (!upload) {
74
      return Response.json(uploadError)
75
    }
76
    const { data } = supabaseClient.storage.from('images').getPublicUrl(upload.path!)
77
    return Response.json(data)
78
  }
79
80
  return Response.json(response)
81
})

Run the function locally#

Run supabase start (see: https://supabase.com/docs/reference/cli/supabase-start)
Start with env: supabase functions serve --env-file supabase/.env
Make an HTTP request:

1
curl -i --location --request POST 'http://127.0.0.1:54321/functions/v1/amazon-bedrock' \
2
    --header 'apikey: <SUPABASE_PUBLISHABLE_KEY>' \
3
    --header 'Content-Type: application/json' \
4
    --data '{"prompt":"A beautiful picture of a bird"}'

Navigate back to your storage bucket. You might have to hit the refresh button to see the uploaded image.

Deploy to your hosted project#

1
supabase link
2
supabase functions deploy amazon-bedrock
3
supabase secrets set --env-file supabase/.env

You've now deployed a serverless function that uses AI to generate and upload images to your Supabase storage bucket.