๐ Complete access to all Anthropic Claude models through Google Vertex AI - including the latest Claude 4 Opus & Sonnet with extended thinking capabilities. โจ FEATURES: - Claude 4 Opus & Sonnet (Latest 2025 models) - Claude 3.7 Sonnet with extended thinking - All Claude 3.5 & 3.0 models included - Vision support for image analysis - Enterprise-grade security & authentication - Multiple auth methods (Service Account, ADC, Environment) - Extended thinking process visualization - Configurable token limits & temperature ๐ง QUICK SETUP: 1. Set your Google Cloud PROJECT_ID in configuration 2. Enable Vertex AI and Claude models in your GCP project 3. Configure authentication (Service Account JSON or Environment) 4. Start chatting with Claude models! ๐ก AUTHENTICATION OPTIONS: - Service Account JSON (paste directly) - GOOGLE_APPLICATION_CREDENTIALS file - Application Default Credentials (ADC) - Environment variables ๐ฏ PERFECT FOR: - Development teams using Google Cloud - Enterprise deployments requiring Vertex AI - Organizations needing advanced Claude capabilities - Projects requiring vision + text processing โก All Claude models work out-of-the-box with proper GCP setup!
Anthropic Claude 4 & All Models via Vertex AI
๐ Complete access to all Anthropic Claude models through Google Vertex AI - including the latest Claude 4 Opus & Sonnet with extended thinking capabilities.
โจ Features
๐ค Complete Model Access
Claude 4 Opus - Most powerful model for complex tasks and coding
Claude 4 Sonnet - Balanced performance and cost efficiency
Claude 3.7 Sonnet - Extended thinking capabilities
Claude 3.5 Sonnet V2 - Advanced reasoning and analysis
Claude 3.5 Haiku - Lightning-fast responses
Claude 3 Opus & Haiku - Previous generation models
๐ง Advanced Capabilities
Extended Thinking - See Claude's reasoning process (Claude 4 & 3.7)
Vision Support - Process images, charts, documents, and screenshots
Long Context - Handle up to 200K tokens of context
Tool Use - Function calling and structured outputs
Multi-modal - Text, image, and document processing
๐ Enterprise Security
Multiple Authentication Methods - Service Account, ADC, Environment Variables
Google Cloud Integration - Native Vertex AI security and compliance
No API Key Management - Uses Google Cloud IAM permissions
Audit Logging - Full request/response logging through Vertex AI
Regional Deployment - Choose your preferred Google Cloud region
Environment Flexibility - Works in any Google Cloud environment
๐ Quick Start
1. Install the Function
Go to Admin Panel > Functions in Open WebUI
Click "+" to add a new function
Search for "Anthropic Claude 4 & All Models via Vertex AI" in the community
Click Install
2. Configure Google Cloud
bash
# Enable Vertex AI API in your Google Cloud project
gcloud services enable aiplatform.googleapis.com
# Enable Claude models in Vertex AI Model Garden
# Go to: https://console.cloud.google.com/vertex-ai/model-garden
# Search for "Claude" and enable the models you want to use
3. Set Up Authentication
Option A: Service Account JSON (Recommended)
Create a service account in Google Cloud Console
Grant it Vertex AI User role
Download the JSON key file
Copy the JSON content and paste it in the SERVICE_ACCOUNT_JSON field
Set your PROJECT_ID (your Google Cloud project ID)
Adjust other settings as needed:
LOCATION: us-east5 (recommended for Claude models)
ENABLE_THINKING: true for Claude 4 thinking capabilities
MAX_OUTPUT_TOKENS: 4096 (or higher for longer responses)
๐ Configuration Reference
Required Settings
Setting
Description
Example
PROJECT_ID
Your Google Cloud Project ID
my-ai-project-123
LOCATION
Vertex AI region
us-east5
Authentication (Choose One)
Method
Setting
Description
Service Account JSON
SERVICE_ACCOUNT_JSON
Paste full JSON content
Environment File
GOOGLE_APPLICATION_CREDENTIALS
Path to JSON file
Default Credentials
None needed
Uses gcloud auth
Optional Settings
Setting
Default
Description
MAX_OUTPUT_TOKENS
4096
Maximum response length
TEMPERATURE
0.7
Response creativity (0.0-1.0)
ENABLE_THINKING
false
Show Claude's reasoning process
THINKING_TOKEN_BUDGET
16000
Tokens for thinking (1024-32000)
๐ฏ Usage Examples
Basic Chat
text
User: Explain quantum computing in simple terms
Claude: [Provides detailed explanation using Claude 4's capabilities]
Extended Thinking (Claude 4)
text
User: Plan a complete marketing strategy for a new SaaS product
๐ค **Thinking Process:**
Let me break this down systematically...
[Shows detailed reasoning process]
**Marketing Strategy:**
[Provides comprehensive strategy based on thinking process]
Vision Analysis
text
User: [Uploads image] What's in this chart and what insights can you provide?
Claude: [Analyzes the image and provides detailed insights]
Code Review
text
User: Review this Python code for security issues
Claude: [Provides detailed code analysis using Claude 4's coding expertise]
๐ง Troubleshooting
Common Issues
โ "Authentication failed"
Verify your service account has Vertex AI User role
Check that your JSON is valid and complete
Ensure your project has Vertex AI API enabled
โ "Invalid Endpoint name"
Verify your PROJECT_ID is correct
Check that the Claude models are enabled in your project
Confirm you're using the correct region (us-east5 recommended)
โ "Model not found"
Go to Vertex AI Model Garden and enable Claude models
Wait a few minutes after enabling for propagation
Check that your region supports Claude models
โ "Quota exceeded"
Check your Vertex AI quotas in Google Cloud Console
Request quota increases if needed
Consider using different regions for better availability
Debug Mode
Enable debug mode by setting DEBUG_MODE = True in the function code:
python
# At the top of the function code
DEBUG_MODE = True
This will show detailed logs including:
Authentication method used
API endpoints called
Request/response payloads
Error details
๐ Supported Regions
Claude models are available in these Vertex AI regions:
us-east5 (Virginia) - Recommended
us-central1 (Iowa)
europe-west4 (Netherlands)
asia-southeast1 (Singapore)
๐ฐ Pricing
Pricing follows Vertex AI's pay-as-you-go model:
Claude 4 Opus: $15 input / $75 output per 1M tokens
Claude 4 Sonnet: $3 input / $15 output per 1M tokens
Claude 3.5 models: $3 input / $15 output per 1M tokens
Claude 3 models: $0.25-$15 per 1M tokens (varies by model)
Prices subject to change. Check Google Cloud Vertex AI pricing for current rates.
๐ Requirements
Dependencies
google-auth - Google Cloud authentication
google-cloud-aiplatform - Vertex AI client library
Permissions
Your service account needs:
Vertex AI User (roles/aiplatform.user)
Service Account Token Creator (if using impersonation)