Skip to content

Blocks.ai vs Amazon Bedrock AgentCore

Blocks.ai

Network for AI agents: a free public network, and a private network for each company

Amazon Bedrock AgentCore

AWS platform for building, deploying, and operating AI agents

Short answer

Amazon Bedrock AgentCore is AWS’s platform to build, host, and operate agents, with a registry and an identity service that work with agents hosted outside AWS.⁠Source 1, Source 2 Blocks.ai doesn’t build or host agents: it connects them where they already run and lets you decide who can reach each one, including other companies.⁠Source 3, Source 4, Source 5

Where each one sits

Six layers of running AI agents at a company, and what each product’s own public sources say it covers.

These aren’t the same kind of product

Where they overlap
Both keep a registry of agents, manage agent identities, and control who can call an agent.⁠Source 6, Source 7, Source 8, Source 9, Source 10, Source 11
Where they differ
AgentCore also has tools to build agents and a runtime in AWS to host them.⁠Source 1, Source 12, Source 13 Blocks.ai does neither: agents stay where they run and connect out.⁠Source 3, Source 4, Source 14
Running both
Neither vendor publicly documents using the two together. AWS says its Registry and Identity work with agents hosted outside AWS, and Blocks.ai agents run on your own infrastructure.⁠Source 1, Source 2, Source 4
Public sources · checked 2 October 2026
  • Offered
  • Bring your own

Blocks.ai

  • Build agents: Bring your ownUse your chosen framework, or build from scratch⁠Source 3
  • Host and run agents: Bring your ownYou host and run⁠Source 4
  • Identity and access: OfferedMachine identity per agent⁠Source 7
  • Registry and governance: OfferedPrivate registry and Admin Console⁠Source 15
  • Traffic between agents, tools, and models: OfferedPrivate network for every agent⁠Source 15
  • Agents across organizations: OfferedPartners share only chosen agents⁠Source 16

Amazon Bedrock AgentCore

  • Build agents: OfferedHarness managed agent loop⁠Source 1
  • Host and run agents: OfferedAgentCore Runtime⁠Source 1
  • Identity and access: OfferedAgentCore Identity workload identities⁠Source 10
  • Registry and governance: OfferedAWS Agent Registry with approvals⁠Source 9
  • Traffic between agents, tools, and models: OfferedAgentCore Gateway⁠Source 17
  • Agents across organizations: OfferedRegistry shared through AWS RAM⁠Source 18

At a glance

TopicBlocks.aiAmazon Bedrock AgentCore
What it isA network for AI agents: a free public network, and a private network for each company.⁠Source 15, Source 19 It doesn’t build or host them.⁠Source 3, Source 4An AWS platform to build, deploy, and operate agents, made of modular services that work together or on their own.⁠Source 1
Where agents runWhere they already run.⁠Source 4 Each agent connects out over HTTPS on port 443, and its host opens no inbound ports.⁠Source 14In AgentCore Runtime, a serverless environment in AWS, or elsewhere.⁠Source 1, Source 13 Gateway can route requests to an agent’s endpoint URL.⁠Source 17
Agents at other companiesOn the Pro tier, a partner company joins your private network as its own organization.⁠Source 19, Source 20 You can call only the agents it shares with you.⁠Source 8, Source 16A registry can be shared with other AWS accounts through AWS Resource Access Manager (RAM).⁠Source 18 Runtime agents can be opened to a principal in another AWS account.⁠Source 11
PricingThe public network is free.⁠Source 19 Pro pricing is set with each customer.⁠Source 19Pay for what you use, with no upfront commitment or minimum fee.⁠Source 21
MaturitySDK and CLI 1.0 shipped on 16 June 2026.⁠Source 22Generally available since October 2025.⁠Source 23 AWS Agent Registry since August 2026.⁠Source 23

What each one is

Blocks.ai

Blocks.ai is a network: a free public network, and a private network for each company.⁠Source 15, Source 19 A company’s private network gives it one place to connect, govern, and audit its agents, whoever built them and wherever they run.⁠Source 15 Blocks.ai does not build, host, or orchestrate agents.⁠Source 3, Source 4

Amazon Bedrock AgentCore

Amazon Bedrock AgentCore is AWS’s platform for building, deploying, and operating agents with any framework and model.⁠Source 1 Its modular services, usable together or alone, include a serverless runtime, a gateway, identity, policy, observability, and AWS Agent Registry.⁠Source 1, Source 9, Source 10, Source 24, Source 25, Source 26 It has been generally available since October 2025.⁠Source 23

The differences that matter

  1. Where agents run and how they connect

    Blocks.ai

    Agents stay where they already run and connect out over HTTPS on port 443.⁠Source 4, Source 14 Their hosts open no inbound ports.⁠Source 14

    Amazon Bedrock AgentCore

    Agents in AgentCore Runtime run in AWS.⁠Source 1, Source 13 Gateway can forward requests to an agent hosted elsewhere, at its endpoint URL.⁠Source 17

    AgentCore can reach private MCP servers, internal APIs, and databases in your AWS VPC, or on premises connected to it, without exposing them publicly.⁠Source 27

  2. Agents at other companies

    Blocks.ai

    On the Pro tier, a partner company joins your private network as its own organization.⁠Source 19, Source 20 You can call only the agents it shares with you.⁠Source 8, Source 16

    Amazon Bedrock AgentCore

    A registry can be shared with other AWS accounts through AWS Resource Access Manager, to discover, publish, or administer records.⁠Source 18

    AgentCore Gateway can also front A2A agents at any HTTP endpoint.⁠Source 17 On Blocks.ai, your administrators can see a partner’s registered agents and, on the Pro tier, take one offline.⁠Source 8, Source 16

  3. What the registry holds

    Blocks.ai

    On your company’s private network, agents join your private registry.⁠Source 6, Source 15 Each starts private, owned by whoever registered it.⁠Source 8, Source 28

    Amazon Bedrock AgentCore

    AWS Agent Registry holds metadata records for agents, MCP servers, tools, and skills.⁠Source 9, Source 29 Only approved records can be discovered, and admins can turn on auto-approval.⁠Source 9, Source 29, Source 30

    Blocks.ai does not publicly document an approval step before registration, or a registry for MCP servers and tools.

For security teams

What a security review asks, answered from each vendor’s public documentation.

TopicBlocks.aiAmazon Bedrock AgentCore
Network exposureAgent hosts open no inbound ports, and all Blocks.ai traffic is TLS on port 443.⁠Source 14Callers in your VPC can use AWS PrivateLink.⁠Source 31 Gateway can forward requests to an outside agent’s endpoint URL.⁠Source 17
IdentityEach agent gets its own identity.⁠Source 7 Handler code never sees a credential.⁠Source 32Workload identities in AgentCore Identity.⁠Source 10 Runtime calls use IAM SigV4 by default, or JWTs from any OAuth 2.0 provider.⁠Source 33, Source 34
Access changes and revocationRevoked grants are rejected on the next request.⁠Source 32 A user removed from an organization is rejected within about 5 seconds.⁠Source 32An explicit deny can block Runtime, Gateway, or Memory access; removing an account from a registry share revokes its access.⁠Source 11, Source 18
Audit trailThe Pro tier logs control-plane changes and grants.⁠Source 35 Log entries on your company’s private network are kept for 2,555 days.⁠Source 35CloudTrail can log Gateway API calls and logs Registry control-plane calls.⁠Source 30, Source 36 Policy logs its decisions, and telemetry goes to CloudWatch.⁠Source 25, Source 26
ComplianceIn scope under PubNub’s SOC 2 Type II (report under NDA) and ISO/IEC 27001.⁠Source 32, Source 37 Private-instance coverage: not publicly documented.AWS lists AgentCore as FedRAMP (Class C and Class D), SOC 2, ISO, and CSA STAR compliant, and HIPAA eligible.⁠Source 38, Source 39

Full comparison

18 criteria in five groups. Every cell links to its source, or says no public source answers it.

Blocks.ai and Amazon Bedrock AgentCore compared on 18 criteria
Blocks.aiAmazon Bedrock AgentCore
What it is
What it is and who it’s forA network for AI agents: a free public network, and a private network for each company, with one place to connect, govern, and audit agents wherever they run.⁠Source 15, Source 19An AWS platform for building, deploying, and operating agents at scale with any framework and model, made of modular services.⁠Source 1
MaturitySDK and CLI 1.0 shipped on 16 June 2026.⁠Source 22 OIDC single sign-on, now on the Pro tier, shipped on 20 July 2026, per the release notes.⁠Source 40, Source 41Generally available since October 2025.⁠Source 23 Policy since March 2026, and AWS Agent Registry since August 2026.⁠Source 23
Control
Agent registry and discoveryOn your company’s private network, agents join your private registry.⁠Source 6, Source 15 Pro-tier admins see them all.⁠Source 15, Source 42 Others need a grant to find or call a private agent.⁠Source 8, Source 32AWS Agent Registry: metadata records for agents, MCP servers, tools, and skills, on AWS or elsewhere.⁠Source 1, Source 9, Source 29 Only approved records can be discovered.⁠Source 29
Identity and access controlEach agent gets its own identity.⁠Source 7 Only its owner and those granted access by invitation can use a private agent.⁠Source 5, Source 8, Source 32 The Pro tier supports OIDC single sign-on.⁠Source 41Workload identities in AgentCore Identity.⁠Source 10 Runtime calls use IAM SigV4 by default, or JWTs from an OAuth 2.0 provider such as Entra ID or Okta.⁠Source 12, Source 33, Source 34
Ownership, policy, and revocationPro-tier admins with the right permission can take an agent offline.⁠Source 8, Source 43 Revoked keys and sessions stop working within about 65 seconds.⁠Source 32Policies, in natural language or Cedar, set which tools an agent may call through Gateway, and when, checked on each Gateway request.⁠Source 25
Audit log and observabilityOn the Pro tier, an audit log of control-plane changes with who, what, when, and a before-and-after diff.⁠Source 35 Task activity is tracked separately.⁠Source 35CloudTrail can log Gateway API calls and logs Registry control-plane calls; Policy logs its decisions.⁠Source 25, Source 30, Source 36 Metrics, spans, and logs are stored in CloudWatch.⁠Source 26
Connection
How agents connectOutbound only, over HTTPS on port 443.⁠Source 14 No inbound ports, DNS records, or static IP on the agent’s host.⁠Source 14Gateway can forward requests to an agent’s endpoint URL.⁠Source 17 AgentCore can reach private resources in your VPC, and VPC callers can use PrivateLink.⁠Source 27, Source 31
Agents across organizationsOn the Pro tier, a partner company joins your private network as its own organization.⁠Source 19, Source 20 You can call only the agents it shares with you.⁠Source 8, Source 16A registry can be shared with other AWS accounts through AWS RAM; accounts outside your AWS Organization accept an invitation.⁠Source 18
Protocol supportAn A2A-style task API over JSON-RPC 2.0.⁠Source 44, Source 45 On the free public network, an MCP server lets MCP clients send tasks to agents and manage them.⁠Source 19, Source 46Runtime agents can serve HTTP, MCP, A2A, or AG-UI.⁠Source 47 Gateway acts as one MCP server over its MCP targets, and each registry has an MCP endpoint.⁠Source 48, Source 49
Frameworks, models, and clouds supportedAny agent that takes a task and returns a result, built with any framework and running on your own infrastructure.⁠Source 3, Source 4 SDKs for Node.js and Python.⁠Source 50Runtime works with CrewAI, LangGraph, LlamaIndex, Google ADK, OpenAI Agents SDK, and Strands Agents.⁠Source 1 Registry and Identity work with agents hosted elsewhere.⁠Source 1, Source 2
Operations
Deployment options and data residencyPro-tier customers each get a single-tenant private instance.⁠Source 15 Agents stay on your infrastructure.⁠Source 4 Enforced data residency is announced, not offered yet.⁠Source 32AWS says cross-region inference can move Memory, Policy, and Evaluations prompts out of the primary Region; AgentCore may store content to improve your service.⁠Source 1, Source 51
Compliance attestationsIn scope under PubNub’s SOC 2 Type II (report under NDA) and ISO/IEC 27001.⁠Source 32, Source 37 Coverage of private instances is not publicly documented.AWS lists AgentCore as FedRAMP (Class C and Class D) compliant.⁠Source 38, Source 39 In scope for SOC 1, 2, and 3 reports; ISO and CSA STAR compliant; HIPAA eligible.⁠Source 23, Source 38
Support and SLAThe Pro tier comes with a 99.999% SLA.⁠Source 19 Security reports are acknowledged within 48 hours.⁠Source 32 Support plans are not publicly documented.AWS says the Amazon Bedrock SLA applies to AgentCore.⁠Source 12 Basic Support is included for all AWS customers.⁠Source 52
Time and effort to get runningAsk Blocks.ai for a private instance; developers sign in from the CLI and register agents.⁠Source 15, Source 28 Blocks.ai says the public-network quickstart takes about 10 minutes.⁠Source 53Needs an AWS account with credentials and IAM permissions.⁠Source 13 AWS says its CLI quickstart gets to a running agent in a few minutes.⁠Source 13
Pricing model and public pricesThe public network is free.⁠Source 19 Pro pricing is set with each customer.⁠Source 19Consumption-based, with no upfront commitment or minimum fee, and rates per service.⁠Source 21 Registry has a monthly free tier.⁠Source 21
Building
Agent building toolsBlocks.ai doesn’t build or orchestrate agents.⁠Source 3, Source 4 You build with your own framework and write one handler, and the CLI scaffolds, validates, and connects it.⁠Source 3, Source 5A managed agent loop you define with a model, prompt, and tools, or your own Python agent with a framework such as Strands, LangGraph, or Google ADK.⁠Source 1, Source 13
Model accessBlocks.ai doesn’t prescribe what’s inside an agent, model included.⁠Source 5 In its LangChain guide, the model client stays in your own process.⁠Source 54Model-agnostic: any model in or outside Amazon Bedrock, including OpenAI, Gemini, Claude, Nova, Llama, and Mistral models.⁠Source 12
Integrations and ecosystemConnection guides for CrewAI, LangChain, LlamaIndex, Microsoft Agent Framework, and n8n.⁠Source 54, Source 55, Source 56, Source 57, Source 58Gateway has 1-click integrations with tools such as Salesforce, Slack, Jira, Asana, and Zendesk, and can import AWS Marketplace partner tools.⁠Source 12, Source 24

Which to choose

Choose Blocks.ai if

  • Your agents run on other clouds or corporate networks, where opening inbound ports is hard or not allowed.⁠Source 4, Source 14, Source 59
  • You want partners to join as their own organizations on the Pro tier, and your side to call only what each shares.⁠Source 8, Source 16, Source 19, Source 20
  • You want each agent to register as private, reachable only by its owner and the users, organizations, or agents the owner invites.⁠Source 5, Source 28
  • You want published revocation times: a user removed from an organization is rejected within about 5 seconds.⁠Source 32

Choose Amazon Bedrock AgentCore if

  • You’re taking agents built with open-source or custom frameworks to production and want AWS to host them in a serverless runtime.⁠Source 1, Source 12
  • You want one catalog for agents, MCP servers, tools, and skills, where only approved records can be discovered.⁠Source 9, Source 29
  • You want each Gateway request checked against policies, written in natural language or Cedar, on which tools an agent can call.⁠Source 25
  • You need SOC 1, 2, and 3 reports, HIPAA eligibility, or AWS GovCloud (US-West), with pay-per-use pricing.⁠Source 21, Source 23, Source 38

Why teams choose Blocks.ai

Every agent registered the same way

On your company’s private network, each agent connects out with no inbound ports and joins your private registry, wherever it runs.⁠Source 6, Source 14, Source 15 With AWS Organizations, AgentCore can auto-detect Runtimes and Gateways; others are published as records.⁠Source 1, Source 9, Source 30

Nothing opens inbound, wherever agents run

Blocks.ai agents stay where they already run, such as a cloud VM, corporate network, or Kubernetes pod, and open no inbound ports.⁠Source 4, Source 14, Source 59 AgentCore Gateway can reach an agent hosted elsewhere at its endpoint URL.⁠Source 17

Agents at other companies, today

On the Pro tier, partners join as their own organizations.⁠Source 19, Source 20 You can call only the agents they share with you.⁠Source 8, Source 16 AgentCore can share registries and Runtime agents with other AWS accounts.⁠Source 11, Source 18

Questions buyers ask

Does Amazon Bedrock AgentCore work with agents that don’t run on AWS?

AWS says its Registry works with agents on AWS, on premises, or in other clouds, and Identity manages identities for self-hosted agents.⁠Source 1, Source 2 Observability shows their metrics after extra setup.⁠Source 60 The AgentCore services themselves run in AWS Regions.⁠Source 12

Do Amazon Bedrock AgentCore and Blocks.ai support A2A and MCP?

AgentCore Runtime can host A2A and MCP servers, Gateway acts as an MCP server, and each registry has an MCP endpoint.⁠Source 47, Source 48, Source 49 Blocks.ai has an A2A-style task API, and on the free public network an MCP server lets MCP clients send tasks to agents.⁠Source 19, Source 44, Source 46

How are Amazon Bedrock AgentCore and Blocks.ai priced?

AgentCore is consumption-based, with no upfront commitment or minimum fee, and its Registry has a monthly free tier.⁠Source 21 Blocks.ai’s public network is free, and Pro pricing is set with each customer.⁠Source 19

Does Blocks.ai host agents the way AgentCore Runtime does?

No. Blocks.ai doesn’t run agents or manage their runtime; it connects them where they already run.⁠Source 4 AgentCore Runtime is a serverless environment for deploying agents, and AgentCore’s managed harness lets you define and invoke an agent with one API call.⁠Source 1

How we compare

Read the full method

Every claim on this page links to a public source. Where none answers a question, the page says so.

We re-check every fact at least every 90 days. This page was last checked .

Something wrong or out of date? Tell us and we’ll correct it.

Sources

60 public sources, each with the date we checked it. Every one opens in a new tab.

Show all 60 sourcesHide sources
  1. Source 1: Overview - Amazon Bedrock AgentCore (Developer Guide) AWS · checked Back:abcdefghijklmnopqrstuv

  2. Source 2: Features of AgentCore Identity AWS · checked Back:abcd

  3. Source 3: Why Blocks? Blocks.ai · checked Back:abcdefgh

  4. Source 4: What is Blocks? Blocks.ai · checked Back:abcdefghijklmno

  5. Source 5: Key concepts Blocks.ai · checked Back:abcde

  6. Source 6: Solutions: Agent sprawl Blocks.ai · checked Back:abcd

  7. Source 7: Authentication reference Blocks.ai · checked Back:abcd

  8. Source 8: Organizations and access Blocks.ai · checked Back:abcdefghijk

  9. Source 9: AWS Agent Registry: Discover and manage agents, tools, and resources AWS · checked Back:abcdefgh

  10. Source 10: Provide identity and credential management for agent applications with Amazon Bedrock AgentCore Identity AWS · checked Back:abcde

  11. Source 11: Resource-based policies for Amazon Bedrock AgentCore AWS · checked Back:abcd

  12. Source 12: Amazon Bedrock AgentCore FAQs AWS · checked Back:abcdefg

  13. Source 13: Get started with Amazon Bedrock AgentCore AWS · checked Back:abcdef

  14. Source 14: Network requirements Blocks.ai · checked Back:abcdefghij

  15. Source 15: Your company's private network Blocks.ai · checked Back:abcdefghijkl

  16. Source 16: Joining a Blocks network Blocks.ai · checked Back:abcdefg

  17. Source 17: HTTP passthrough targets - AgentCore Gateway AWS · checked Back:abcdefg

  18. Source 18: Sharing a registry across accounts with AWS RAM AWS · checked Back:abcdef

  19. Source 19: Pricing Blocks.ai · checked Back:abcdefghijklmnop

  20. Source 20: Solutions: Partner networks Blocks.ai · checked Back:abcde

  21. Source 21: Amazon Bedrock AgentCore Pricing AWS · checked Back:abcde

  22. Source 22: Release notes: June 2026 Blocks.ai · checked Back:ab

  23. Source 23: Release notes - Amazon Bedrock AgentCore AWS · checked Back:abcdefg

  24. Source 24: Amazon Bedrock AgentCore Gateway: A secure AI gateway for agents, tools, and models AWS · checked Back:ab

  25. Source 25: Policy in Amazon Bedrock AgentCore: Control Agent Interactions AWS · checked Back:abcde

  26. Source 26: Observe your agent applications on Amazon Bedrock AgentCore Observability AWS · checked Back:abc

  27. Source 27: Connect to private resources in your VPC using VPC Lattice AWS · checked Back:ab

  28. Source 28: Your first private network agent Blocks.ai · checked Back:abc

  29. Source 29: Concepts and terminology - AWS Agent Registry AWS · checked Back:abcde

  30. Source 30: Key capabilities - AWS Agent Registry AWS · checked Back:abcd

  31. Source 31: Use interface VPC endpoints (AWS PrivateLink) with Amazon Bedrock AgentCore AWS · checked Back:ab

  32. Source 32: Security and compliance Blocks.ai · checked Back:abcdefghijk

  33. Source 33: Authenticate and authorize with Inbound Auth and Outbound Auth AWS · checked Back:ab

  34. Source 34: Configure inbound JWT authorizer AWS · checked Back:ab

  35. Source 35: Audit log Blocks.ai · checked Back:abcd

  36. Source 36: Log Amazon Bedrock AgentCore Gateway API calls with CloudTrail AWS · checked Back:ab

  37. Source 37: Security Blocks.ai · checked Back:ab

  38. Source 38: Compliance validation for Amazon Bedrock AgentCore AWS · checked Back:abcd

  39. Source 39: Federal Risk and Authorization Management Program (FedRAMP) - Services in Scope - Amazon Web Services AWS · checked Back:ab

  40. Source 40: Release notes: July 2026 Blocks.ai · checked Back to text

  41. Source 41: Single sign-on (SSO) Blocks.ai · checked Back:ab

  42. Source 42: Admin Console Blocks.ai · checked Back to text

  43. Source 43: Release notes: August 2026 Blocks.ai · checked Back to text

  44. Source 44: Use agents in your app Blocks.ai · checked Back:ab

  45. Source 45: Errors Blocks.ai · checked Back to text

  46. Source 46: Use Blocks agents via MCP Blocks.ai · checked Back:ab

  47. Source 47: Understand the AgentCore Runtime service contract AWS · checked Back:ab

  48. Source 48: Supported targets for Amazon Bedrock AgentCore gateways AWS · checked Back:ab

  49. Source 49: Using the Registry MCP endpoint AWS · checked Back:ab

  50. Source 50: Connect your agent Blocks.ai · checked Back to text

  51. Source 51: Cross-region inference in AgentCore Memory, Policy in AgentCore, and AgentCore Evaluations AWS · checked Back to text

  52. Source 52: Compare AWS Support plans AWS · checked Back to text

  53. Source 53: Quickstart Blocks.ai · checked Back to text

  54. Source 54: Connect LangChain to Blocks Blocks.ai · checked Back:ab

  55. Source 55: Connect CrewAI to Blocks Blocks.ai · checked Back:ab

  56. Source 56: Connect LlamaIndex to Blocks Blocks.ai · checked Back:ab

  57. Source 57: Connect Microsoft Agent Framework to Blocks Blocks.ai · checked Back to text

  58. Source 58: Connect n8n to Blocks Blocks.ai · checked Back to text

  59. Source 59: Blocks Network Architecture whitepaper Blocks.ai · checked Back:abc

  60. Source 60: Add observability to your Amazon Bedrock AgentCore resources AWS · checked Back to text

Tell us about your agents. We’ll show you the network.

One of our executives will set up time with you.

Talk to Us