← Back to all announcements
★★★★☆ 30/04/2026

Amazon Bedrock AgentCore Identity now supports On-Behalf-Of (OBO) token exchange

Agents can now securely access downstream resources on behalf of users without requiring separate permission flows for each resource.

View original announcement →

Visual Summary

graph TD A{{AgentCore Identity OBO Exchange}}:::announced B((End User)):::external C(Amazon Bedrock AgentCore):::compute D([User Access Token]):::feature E([Scoped OBO Token]):::feature F((Protected Resource)):::external G([Least-Privilege Access]):::feature H(AI Agent):::compute I([Identity Chain Audit]):::feature B ==>|"authenticates"| D D ==>|"presented by"| H H ==>|"requests exchange"| A A -->|"validates & issues"| E E -->|"carries user + agent"| G E ==>|"grants access"| F A -->|"orchestrated within"| C A -.->|"enables"| I I -.->|"traces back to"| B classDef announced fill:#ff9900,stroke:#ec7211,color:#fff,font-weight:bold classDef compute fill:#e3f2fd,stroke:#1565c0,color:#1565c0 classDef storage fill:#e8f5e9,stroke:#2e7d32,color:#2e7d32 classDef feature fill:#fff3e0,stroke:#e65100,color:#e65100 classDef external fill:#f5f5f5,stroke:#616161,color:#616161

What's New

Amazon Bedrock AgentCore Identity now supports On-Behalf-Of (OBO) token exchange, a generally available capability that enables AI agents to securely access protected downstream resources while preserving the identity of the authenticated end user. Developers can exchange an existing access token for a new, scoped-down token that carries both the original user identity and the agent identity, targeted specifically to the outbound resource being accessed. This eliminates the need for users to complete separate consent flows for each protected resource an agent needs to interact with.

How It Works

  • OBO token exchange follows a delegated authorization pattern rooted in OAuth 2.0 token exchange semantics (RFC 8693).
  • When an authenticated user initiates an agent workflow, the agent receives the user's access token and presents it to Amazon Bedrock AgentCore Identity along with a request for a new token scoped to a specific downstream protected resource.
  • AgentCore Identity validates the incoming token, verifies the agent's identity, and issues a new access token that encodes both the original user principal and the agent principal.
  • This derived token is narrowly scoped — granting just-in-time, least-privilege access — and is targeted exclusively to the intended outbound resource, meaning it cannot be reused across other services.
  • The process is transparent to the end user and requires no additional consent prompts after the initial authorization.

Why It's Important

  • This capability is significant because it directly addresses one of the most common friction points in agentic AI architectures: managing delegated authorization across multiple protected resources without degrading the user experience.
  • In enterprise and consumer applications where agents must orchestrate calls to APIs, databases, or SaaS services on a user's behalf, the traditional approach of triggering separate OAuth consent flows per resource creates both UX friction and security surface area.
  • OBO token exchange centralizes and automates this delegation, enabling developers to build more capable, multi-step agents while adhering to least-privilege security principles.
  • It also ensures auditability, since the derived token carries the full identity chain (user + agent), making it possible to trace actions back to both the originating user and the acting agent.

How It's Different

  • Previously, developers building agents on AWS that needed to act on behalf of users had to manually orchestrate separate consent flows for each protected resource, often relying on custom token management logic, stored refresh tokens, or workarounds that either over-provisioned permissions or required repeated user interaction.
  • There was no native, platform-managed mechanism within Bedrock to perform delegated token exchange that preserved user identity across the agent's downstream calls.
  • OBO token exchange replaces these ad hoc approaches with a standardized, platform-enforced delegation model.
  • Compared to service-to-service IAM role assumption (which carries only the agent/service identity and loses the user context), OBO tokens maintain the full identity chain, which is critical for user-scoped authorization policies and audit logging in downstream systems.

When to Prefer It

  • OBO token exchange is the right choice whenever an AI agent must access user-owned or user-scoped protected resources — such as a user's calendar, email, CRM records, or personal cloud storage — where the downstream system enforces user-level authorization policies.
  • It is particularly well-suited for enterprise agentic workflows that span multiple APIs or microservices, where re-prompting users for consent at each step would be impractical.
  • Prefer OBO over static service credentials or broad IAM roles when auditability of user-level actions is required, when downstream systems enforce per-user access controls, or when compliance requirements mandate that delegated actions be traceable to the originating human identity.
  • It is less appropriate for purely system-to-system workflows where no human user context exists, in which case standard IAM role-based access remains the better fit.

Availability

  • Amazon Bedrock AgentCore Identity OBO token exchange is generally available (GA) as of April 30, 2026.
  • It is supported across 14 AWS Regions: US East (N.
  • Virginia), US East (Ohio), US West (Oregon), Canada (Central), Asia Pacific (Mumbai), Asia Pacific (Seoul), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Ireland), Europe (London), Europe (Paris), and Europe (Stockholm).
  • No preview or limited-access restrictions are noted in the announcement.
  • Specific service quotas, supported identity provider integrations, and any token lifetime or scope limitations are not detailed in the announcement and should be reviewed in the Amazon Bedrock AgentCore Identity documentation before production adoption.

Tags

Servicesbedrock-agentcore
Typega-launchnew-feature
Conceptsagentic-aigenai
Use Casesenterprisedeveloper-tools
GeographyAMERICASAPJEMEA

AI Radar AWS

AWS AI/ML news — curated, researched, explained

An automated intelligence platform that curates, researches, and analyzes AWS AI/ML/GenAI announcements daily. Every report is backed by real research — the system reads linked blog posts and documentation to provide accurate, in-depth analysis.

How Each Report Is Generated

  1. Collection — Daily monitoring of the AWS "What's New" RSS feed
  2. Filtering — AI-powered relevance detection for AI/ML/GenAI topics
  3. Taxonomy Tagging — LLM-based classification across 6 dimensions
  4. Importance Scoring — Point-based system with tag bonuses (1-5 stars)
  5. Research Phase — Follows links to blog posts and documentation
  6. Report Generation — Claude Sonnet produces structured 6-section analysis
  7. Visual Summary — Claude Opus generates Mermaid diagrams for key items
  8. Publishing — Static website rebuilt and deployed via CloudFront

Features

  • Faceted filtering by service, type, concept, and more
  • Multi-dimensional taxonomy with 80+ tags across 6 dimensions
  • Geographic availability badges (Global, APJ, EMEA, AMER) with filtering
  • Timeline visualization of announcement volume
  • PDF export for offline reading
  • Mermaid visual summaries for key announcements
  • Daily automated updates — no manual curation
What makes this different: Each report involves a dedicated research phase where the system reads linked blog posts and AWS documentation pages. This produces analysis that goes beyond the original announcement text.

Technology

Built with Python, AWS Lambda, Amazon Bedrock (Claude Sonnet 4.6, Opus 4.6, Haiku 4.5), S3, CloudFront, WAF, EventBridge, and CDK.

Open Source

This project is open source. Fork it, customize it for your needs, and deploy your own instance.
📦 github.com/bbonik/ai-radar-aws

How Importance Scoring Works

Each announcement receives a point score based on multiple factors. The total score maps to a 1-5 star rating:

1★ < 2 pts 2★ ≥ 2 pts 3★ ≥ 3.5 pts 4★ ≥ 5 pts 5★ ≥ 6.5 pts

Point Breakdown

FactorPointsWhen
Core AI service (Bedrock, AgentCore, SageMaker AI)+4Service named in title
Key AI service (SageMaker, Kiro, QuickSight)+2Service named in title
Other AI-related service+1Default
Blog post link+3Link to aws.amazon.com/blogs/
GitHub samples link+2Link to github.com/aws*
Documentation link+1Link to docs.aws.amazon.com/
New model+1.5Tagged as "new-model"
New service+1Tagged as "new-service"
New feature+0.5Tagged as "new-feature"
Anthropic / OpenAI provider+2Provider explicitly mentioned
Instance / notebook announcement-2Hardware/capacity, not feature
Performance / pricing / security-0.5Incremental updates
Region expansion to APJ+1Expands to Asia Pacific
Region expansion (non-APJ only)-1.5Only expands to other regions

Geographic Relevance Badges

Each announcement card shows a small badge indicating whether the feature is available in your region:

🌐 Global Available in all regions
🌏 APJ Asia Pacific
🌍 EMEA Europe / Middle East / Africa
🌎 AMER Americas (US, Canada, South America)
No badge Geography unknown
How geography is detected: The system detects ALL geographies mentioned in each announcement. If the text mentions specific regions (Tokyo, Frankfurt, Oregon, etc.), the corresponding geography badges are shown. If it says "all regions" or is a new feature with no region specified, it gets the Global badge. Geography is also filterable — click a geo chip to see only announcements available in that region.