Amazon Bedrock AI Model Configuration in ConvyAI - Anthropic Claude
Overview
Amazon Bedrock is a fully managed AWS service that provides access to foundation models from leading AI providers via API - including Anthropic's Claude family. ConvyAI integrates with Bedrock to use Claude models for intent classification, entity extraction, and generative responses.
Prerequisites
- AWS credentials (access key + secret key) with Bedrock permissions.
- Credentials configured in Generative AI Settings.
Create a Bedrock AI Model
Navigate to Generative AI → AI Models and click New Model → Bedrock.
Configuration options
| Field | Description |
|---|---|
| Name | A unique name to identify this AI Model. |
| Access Key | AWS access key with Bedrock permissions. |
| Secret Key | AWS secret key. |
| Region | AWS region where Bedrock is available (e.g., us-east-1). |
Advanced parameters
| Parameter | Default | Description |
|---|---|---|
| Completion model | - | The Anthropic Claude model to invoke, as a free-text field. Enter the model identifier as exposed by Amazon Bedrock. ConvyAI does not check that the model exists or is enabled on your AWS account. |
| Prompt version | v3 | System prompt version applied when using the model in Knowledge Base nodes. |
| Embedding model | cohere.embed-multilingual-v3 | The model used for embedding during Knowledge Base indexing and retrieval. |
| Max token completion | 1000 | Maximum number of tokens in the generated response. |
| Temperature | - | Controls randomness of outputs. Lower values produce more deterministic responses. |
| Top P | - | Nucleus sampling parameter. Limits token selection to the top probability mass. |
| Top K | 250 | Limits token selection to the top-K most probable tokens at each generation step. |
| Extended thinking budget | - | The token budget for extended thinking, as a whole number. Leave empty to use the provider default, enter 0 to turn reasoning off, or -1 to let the model decide dynamically. The value applies to both Knowledge Base and AI Agent nodes, so you configure it once. |
| Number of extracted chunks | 3 | How many Knowledge Base chunks are retrieved and passed to the model as context during RAG. |
Optional parameters start empty on new AI Models. An empty field means the parameter is not sent and the provider default applies; any value you enter is sent as is, 0 included. See Optional parameters.
ConvyAI does not disable or restrict any parameter based on the model you enter. If the model does not support a parameter you set (for example Temperature or Top P on some Claude models, or extended thinking), Amazon Bedrock rejects the call. Use the test console to verify the configuration before deploying. Reasoning content never appears in the response returned to the end user.
Intents, entities, and testing
Configuring intents and entities and testing the model are features shared across all AI Model providers. See Intents, Entities & Testing for full documentation.