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 Zuvo 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 Zuvo project, create a .env file in the supabase directory with the following contents:

AWS_DEFAULT_REGION="<your_region>"
AWS_ACCESS_KEY_ID="<replace_your_own_credentials>"
AWS_SECRET_ACCESS_KEY="<replace_your_own_credentials>"
AWS_SESSION_TOKEN="<replace_your_own_credentials>"

# Mocked config files
AWS_SHARED_CREDENTIALS_FILE="./aws/credentials"
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:

supabase functions new amazon-bedrock

And add the code to the index.ts file:

// We need to mock the file system for the AWS SDK to work.
import { prepareVirtualFile } from 'https://deno.land/x/mock_file@v1.1.2/mod.ts'
import { BedrockRuntimeClient, InvokeModelCommand } from 'npm:@aws-sdk/client-bedrock-runtime@^3'
import { withZuvo } from 'npm:@supabase/server@^1'
import { decode } from 'npm:base64-arraybuffer@^1'

console.log('Hello from Amazon Bedrock!')

// Called with a publishable key on the `apikey` header. Deploy with `verify_jwt = false`.
export default {
  fetch: withZuvo({ auth: 'publishable' }, async (req, ctx) => {
    prepareVirtualFile('./aws/config')
    prepareVirtualFile('./aws/credentials')

    const client = new BedrockRuntimeClient({
      region: Deno.env.get('AWS_DEFAULT_REGION') ?? 'us-west-2',
      credentials: {
        accessKeyId: Deno.env.get('AWS_ACCESS_KEY_ID') ?? '',
        secretAccessKey: Deno.env.get('AWS_SECRET_ACCESS_KEY') ?? '',
        sessionToken: Deno.env.get('AWS_SESSION_TOKEN') ?? '',
      },
    })

    const { prompt, seed } = await req.json()
    console.log(prompt)
    const input = {
      contentType: 'application/json',
      accept: '*/*',
      modelId: 'amazon.titan-image-generator-v1',
      body: JSON.stringify({
        taskType: 'TEXT_IMAGE',
        textToImageParams: { text: prompt },
        imageGenerationConfig: {
          numberOfImages: 1,
          quality: 'standard',
          cfgScale: 8.0,
          height: 512,
          width: 512,
          seed: seed ?? 0,
        },
      }),
    }

    const command = new InvokeModelCommand(input)
    const response = await client.send(command)
    console.log(response)

    if (response.$metadata.httpStatusCode === 200) {
      const { body, $metadata } = response

      const textDecoder = new TextDecoder('utf-8')
      const jsonString = textDecoder.decode(body.buffer)
      const parsedData = JSON.parse(jsonString)
      console.log(parsedData)
      const image = parsedData.images[0]

      const { data: upload, error: uploadError } = await ctx.supabase.storage
        .from('images')
        .upload(`${$metadata.requestId ?? ''}.png`, decode(image), {
          contentType: 'image/png',
          cacheControl: '3600',
          upsert: false,
        })
      if (!upload) {
        return Response.json({ error: uploadError?.message ?? 'Upload failed' }, { status: 500 })
      }
      const { data } = ctx.supabase.storage.from('images').getPublicUrl(upload.path!)
      return Response.json(data)
    }

    return Response.json(response)
  }),
}

Run the function locally

1. Run supabase start (see: https://docs.zuvodev.com/reference/cli/supabase-start)
2. Start with env: supabase functions serve --no-verify-jwt --env-file supabase/.env
3. Make an HTTP request:

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

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

Deploy to your hosted project

supabase link
supabase functions deploy amazon-bedrock --no-verify-jwt
supabase secrets set --env-file supabase/.env

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