Build Your Own AI Assistant with Molt Bot (Clawdbot) and Amazon Nova: A Complete Setup Guide

Learn how to deploy Molt Bot (previously known as Clawdbot) — a powerful, open-source AI assistant — using Amazon Nova foundation models on Amazon Bedrock. This step-by-step guide covers creating AWS access keys, configuring AWS authentication on Molt Bot, and running your own AI assistant powered by AWS-native models.

Yung Jun

Published Jan 27, 2026

Last modified Jan 27, 2026

Amazon Nova is AWS's state-of-the-art foundation models designed for a wide range of AI tasks. From the lightweight Nova Micro for text-only workloads to Nova Pro and Nova Premier for multimodal understanding across text, images, and video — Nova models offer compelling performance at competitive price points.

Molt Bot (previously known as Clawdbot) is an open-source AI assistant that integrates seamlessly with Amazon Bedrock. By combining Molt Bot with Amazon Nova, you get a fully-featured AI assistant powered by AWS-native models.

This guide walks you through setting up Molt Bot with Amazon Nova models using AWS access keys or AWS profile.

Prerequisites

Before you begin, ensure you have the following:

An AWS account
Node.js v24 or later installed on your machine

Amazon Nova 2 Model Overview

Amazon Nova offers two generations of models. Nova 2 is the latest release with enhanced reasoning, extended thinking, and built-in tools. Nova models remain available for existing workloads.

Nova 2 Models (Latest)

Model	Model ID	Input	Output	Context	Best For
Nova 2 Lite	amazon.nova-2-lite-v1:0	Text, Image, Video, Documents	Text	1M tokens	High-volume agentic workflows, extended thinking, cost-efficient multimodal

Nova 2 Key Features:

Extended thinking - Advanced reasoning with three-level intensity control for complex multi-step problems
Built-in tools - Web grounding for real-time information and code interpreter for Python execution
AI agent optimization - Improved tool use and context maintenance for agentic workflows
Enhanced document understanding - Better processing of tables, charts, and multi-page documents

Step 1: Create IAM User

Navigate to IAM in the AWS Console
Select Users from the left navigation pane
Select Create Users and fill in the User name, for example "MoltBot"
Click Attach policies directly
Under Permissions polices, select AmazonBedrockFullAccess
Click Next
Click Create user

Step 2: Create Access Keys

Go to the Security credentials tab
Under Access keys, click Create access key
Select Command Line Interface (CLI) as the use case, tick the checkbox and click Next
Click Create access key
Download the CSV file or copy both the Access Key ID and Secret Access Key

Important: You will not be able to view the secret key again after this step.

Step 3: Install & Onboard Molt Bot (Clawdbot)

Verify if you have Node.js installed:

1
node --version

Install Molt Bot (Clawdbot):

1
npm install -g clawdbot@latest

Important: As of 27-Jan-2026, the npm package is still named as clawdbot, hence use clawdbot for all command until the npm package name is updated to moltbot.

Verify the installation:

1
clawdbot --version

Run Onboarding:

1
clawdbot onboard --install-daemon

During the onboarding wizard:

Accept the security notice
Select Manual onboarding mode
Select Local gateway (this machine)
Set your workspace directory
Skip model/auth provider for now
Filter models by provider: select amazon-bedrock
Choose your default model (e.g., amazon-bedrock/amazon.nova-lite-v1:0)
Accept default gateway settings (port 18789, loopback)
Complete the remaining prompts
Select Do this later when prompted if you want to hatch your bot

Step 4: Configure AWS Credentials

Molt Bot (Clawdbot) reads environment variables from ~/.clawdbot/.env. Create this file with your AWS credentials using either Option A or Option B:

Option A: Using AWS Access Keys (Recommended for simplicity)

1
2
3
4
5
6
cat > ~/.clawdbot/.env << 'EOF'
# AWS Credentials for Bedrock
AWS_ACCESS_KEY_ID=your-access-key-id-here
AWS_SECRET_ACCESS_KEY=your-secret-access-key-here
AWS_REGION=us-east-1
EOF

Replace your-access-key-id-here and your-secret-access-key-here with your actual AWS credentials.

Option B: Using AWS Profile

If you have AWS CLI configured with profiles (~/.aws/credentials):

1
2
3
4
5
cat > ~/.clawdbot/.env << 'EOF'
# AWS Credentials for Bedrock
AWS_PROFILE=your-profile-name
AWS_REGION=us-east-1
EOF

Replace your-profile-name with your actual AWS profile name.

Step 5: Add Nova 2 Lite to clawdbot.json

nano ~/.clawdbot/clawdbot.json and add the models section:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
{

  "models": {

    "providers": {

      "amazon-bedrock": {

        "baseUrl": "https://bedrock-runtime.us-east-1.amazonaws.com",

        "api": "bedrock-converse-stream",

        "auth": "aws-sdk",

        "models": [

          {

            "id": "us.amazon.nova-2-lite-v1:0",

            "name": "Amazon Nova 2 Lite (US Inference Profile)",

            "reasoning": false,

            "input": ["text", "image"],

            "cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 },

            "contextWindow": 300000,

            "maxTokens": 8192

          }

        ]

      }

    }

  }

}

Important: Merge this into your existing clawdbot.json, don't replace the whole file. It is going to look similar to example below:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
{
  "messages": {
    "ackReactionScope": "group-mentions"
  },
  "models": {
    "providers": {
      "amazon-bedrock": {
        "baseUrl": "https://bedrock-runtime.us-east-1.amazonaws.com",
        "api": "bedrock-converse-stream",
        "auth": "aws-sdk",
        "models": [
          {
            "id": "us.amazon.nova-2-lite-v1:0",
            "name": "Amazon Nova 2 Lite (US Inference Profile)",
            "reasoning": false,
            "input": ["text", "image"],
            "cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 },
            "contextWindow": 300000,
            "maxTokens": 8192
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "maxConcurrent": 4,
      "subagents": {
        "maxConcurrent": 8
      },
      "compaction": {
        "mode": "safeguard"
      },
      "workspace": "/Users/yungjunyong/clawd",
      "model": {
        "primary": "amazon-bedrock/us.amazon.nova-2-lite-v1:0"
      },
      "models": {
        "amazon-bedrock/us.amazon.nova-2-lite-v1:0": {}
      }
    }
  },
  "gateway": {
    "mode": "local",
    "auth": {
      "mode": "token",
      "token": "undefined"
    },
    "port": 18789,
    "bind": "loopback",
    "tailscale": {
      "mode": "off",
      "resetOnExit": false
    }
  },
  "wizard": {
    "lastRunAt": "2026-01-27T18:09:07.983Z",
    "lastRunVersion": "2026.1.24-3",
    "lastRunCommand": "onboard",
    "lastRunMode": "local"
  },
  "meta": {
    "lastTouchedVersion": "2026.1.24-3",
    "lastTouchedAt": "2026-01-27T18:09:07.992Z"
  }
}

Step 6: Restart the Gateway

1
clawdbot daemon restart

Step 7: Verify Setup

Check that models are available and authenticated:

1
clawdbot models list

You should see output like:

1
2
Model                                      Input      Ctx      Local Auth  Tags
amazon-bedrock/us.amazon.nova-2-lite-v1:0     text+image 31k      no    yes   default,configured

The Auth: yes column confirms your AWS credentials are working.

To see all available Bedrock models:

1
clawdbot models list --all

Step 8: Start Using Clawdbot

Launch the terminal UI:

1
clawdbot tui

Or send a message directly:

1
clawdbot agent --message "which AI model powered you"

Changing the Default Model

To switch to a different model (e.g., Claude Opus 3):

1
clawdbot config set agents.defaults.model.primary "amazon-bedrock/anthropic.claude-3-opus-20240229-v1:0"

Conclusion

You now have a fully functional AI assistant powered by Amazon Nova. Molt Bot with Amazon Nova gives you:

AWS-native AI - built and optimized by AWS for Bedrock
Data sovereignty - all inference happens within your AWS account
Flexible pricing - choose the Amazon Nova model that fits your budget
Multimodal capabilities - process text, images, and video with Nova Lite, Pro, or Premier
Enterprise-ready - leverage IAM, VPC, and CloudTrail for governance

From here, you can extend Molt Bot with messaging integrations (Telegram, Slack, Discord), scheduled tasks, memory persistence, and custom skills. Check out the full documentation at docs.molt.bot for more advanced configurations.

# agentic-ai # ai # agents # ai-agents # amazon-bedrock

Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.

Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article

Builder Center

Build Your Own AI Assistant with Molt Bot (Clawdbot) and Amazon Nova: A Complete Setup Guide

Yung Jun

Prerequisites

Amazon Nova 2 Model Overview

Nova 2 Models (Latest)

Step 1: Create IAM User

Step 2: Create Access Keys

Step 3: Install & Onboard Molt Bot (Clawdbot)

Step 4: Configure AWS Credentials

Option A: Using AWS Access Keys (Recommended for simplicity)

Option B: Using AWS Profile

Step 5: Add Nova 2 Lite to clawdbot.json

Step 6: Restart the Gateway

Step 7: Verify Setup

Step 8: Start Using Clawdbot

Changing the Default Model

Conclusion