AI Frameworks · aws-samples

bedrock-chat

Bedrock Chat is an AWS-native chatbot platform built on Amazon Bedrock that supports multi-tenant chat, custom bots with knowledge bases (RAG), bot sharing, and agent-based task automation. Deployed via CloudFormation with optional security controls, it targets teams already invested in AWS infrastructure.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/aws-samples/bedrock-chat

1.3k

GitHub stars

532

Forks

TypeScript

Primary language

MIT-0

License (Requires review (not clearly OSI))

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	aws-samples/bedrock-chat
Owner	aws-samples
Primary language	TypeScript
License	MIT-0 — Requires review (not clearly OSI)
Stars	1.3k
Forks	532
Open issues	129
Latest release	v3.17.0 (2026-06-16)
Last updated	2026-06-23
Source	https://github.com/aws-samples/bedrock-chat

What bedrock-chat is

Full-stack TypeScript/Python application using AWS Bedrock LLMs, OpenSearch Serverless for vector storage, Lambda for compute, DynamoDB for state, Cognito for auth, and WebSocket streaming. Supports multi-tenant knowledge base mode to work around Bedrock's 100 KB account limit. Includes CDK infrastructure-as-code deployment with optional Lambda SnapStart, IP/domain filtering, and API publishing.

Quickstart

Get the bedrock-chat source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/aws-samples/bedrock-chat.git2cd bedrock-chat3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Internal Enterprise AI Assistant with Knowledge Base

Organizations can rapidly deploy a conversational AI platform where employees create custom bots with proprietary knowledge (documents, policies, FAQs) shared via internal bot store. Multi-tenant KB mode allows scaling beyond Bedrock's 100 KB per-account limit.

AWS-Native GenAI Product Feature

Teams already using AWS can embed a managed chatbot platform into their product without vendor lock-in beyond Bedrock. Bot personalization, fine-grained permission controls, and admin analytics support multi-user product scenarios.

Automated Task Execution with Agent Functionality

Complex workflows can be automated via the Agent feature, which breaks down user requests into multi-step tasks and integrates external tools. Suitable for IT automation, customer support escalation, or business process orchestration.

Implementation considerations

Enable model access in Bedrock console before deployment; bin.sh deployment supports optional parameters for security (IP filtering, email domain whitelisting, self-registration controls).
V3 migration is breaking: V2 bots become unusable without following the migration guide. Plan upgrade windows and test in non-production first.
Multi-tenant knowledge base mode (default for new bots) shares a single KB with bot-ID filtering; migrate legacy bots to avoid hitting the 100 KB per-account limit.
Cognito user pool administration is required to grant 'CreatingBotAllowed' group membership for users who should author custom bots; default is restricted.
OpenSearch Serverless availability gates deployment to specific regions (us-east-1, us-east-2, us-west-1/2, ap-south-1, ap-northeast-1/2, ap-southeast-1/2, ca-central-1, eu-central-1, eu-west-1/2, eu-south-2, eu-north-1, sa-east-1).

When to avoid it — and what to weigh

Multi-Cloud or Vendor-Neutral Strategy Required — Bedrock Chat is tightly coupled to AWS Bedrock and OpenSearch Serverless. If your architecture must support Azure OpenAI, Google Vertex AI, or on-premise models, this solution will require significant rework.
Minimal Cloud Infrastructure or Serverless Preference — Deployment requires CloudFormation, VPC networking, Cognito user pools, DynamoDB, OpenSearch Serverless, and Lambda. Cost and complexity scale with usage; unsuitable for teams wanting a simple, fully managed SaaS chatbot.
Compliance or Data Residency Outside AWS Regions — Data flows through AWS infrastructure (Bedrock, OpenSearch, DynamoDB). If your jurisdiction prohibits cloud processing or requires on-premise isolation, this will not fit. Supported regions are limited to specific AWS geographies.
Zero AWS Expertise or Limited DevOps Resources — While bin.sh automates CloudFormation, ongoing management (scaling, security patching, multi-region failover, troubleshooting) requires AWS operational knowledge. Not suitable for non-technical teams or those without AWS support.

License & commercial use

MIT-0 (MIT No Attribution) license: permissive OSI-approved license allowing commercial use, modification, and redistribution with no attribution required.

MIT-0 permits commercial use without restriction. However, you remain responsible for Amazon Bedrock licensing (pay-per-token model with separate pricing), AWS infrastructure costs (OpenSearch Serverless, Lambda, DynamoDB, Cognito), and any data privacy obligations. Verify Bedrock terms of service for your use case (e.g., custom model fine-tuning, data retention policies). No proprietary support agreement is implied by this open-source license.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Strong
License clarity	Clear
Deployment complexity	Moderate
DEV.co fit	Good
Assessment confidence	High

Security considerations

Cognito auth is required; fine-grained permission controls for bot creation are available via user groups. Optional deployment parameters include IP range filtering (IPv4/IPv6), email domain whitelisting, and self-registration disable. Lambda SnapStart can be enabled. Data is processed by AWS Bedrock and stored in OpenSearch Serverless and DynamoDB; review Bedrock's data handling and your region's compliance posture. No mention of encryption at rest, audit logging, or vulnerability disclosure process in provided data. Requires review of Bedrock's SOC 2, ISO 27001, and HIPAA eligibility.

Alternatives to consider

LangChain + OpenAI API (Self-Hosted or Vercel)

Vendor-neutral, supports multiple LLM providers, lower AWS lock-in. Requires more engineering effort for knowledge base, multi-tenancy, and agent orchestration but offers flexibility and portability.

Azure OpenAI + Semantic Kernel or Copilot Studio

If your organization standardizes on Azure, Copilot Studio offers native multi-tenant bots with knowledge integration and low-code bot building. Avoids AWS infrastructure.

Anthropic Claude API + Custom App (Flask/Django/FastAPI)

Lightweight, minimal cloud footprint, and Claude model quality. Requires building chat UI, auth, knowledge indexing, and bot store yourself. Better for small-scale deployments or multi-cloud strategies.

Software development agency

Build on bedrock-chat with DEV.co software developers

Bedrock Chat provides a production-ready, open-source foundation for enterprise conversational AI. Deploy with one command, customize with your knowledge, and scale across your organization. Ideal for teams already on AWS.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

aichat serverless-chat-langchainjs ChatALL

Related on DEV.co

Explore the category and the services that help you build with it.

AI Frameworks AI Development Cloud Deployment Enterprise Software Talk to DEV.co

bedrock-chat FAQ

Can I use Bedrock Chat without AWS?

No. It requires an AWS account with Bedrock and OpenSearch Serverless access. You cannot self-host it without AWS infrastructure.

What happens if I upgrade from V2 to V3?

V3 is a breaking change: bots from V2 become unusable without following the migration guide. Plan a maintenance window and test in a non-production environment first.

How many knowledge bases can I create?

Bedrock limits new accounts to 100 KBs per AWS account. Bedrock Chat works around this via multi-tenant mode, where a single KB is shared across multiple bots with bot-ID filtering. Existing V2 bots should be migrated to multi-tenant mode.

Can non-technical users create custom bots?

Yes, the UI allows users to upload documents and set custom instructions. However, only users in the Cognito 'CreatingBotAllowed' group can author bots; admins must grant this permission.

Custom software development services

From first prototype to production, DEV.co delivers software development services around tools like bedrock-chat. Our software development agency staffs experienced software developers and web developers for custom software development, web development, integrations, and ongoing support across ai frameworks and beyond.

Ready to Build an AI-Powered Chatbot on AWS?

Talk to DEV.co Schedule a Consultation