AWS lists 23 open source generative AI assets from its Innovation Center
AWS published a catalog of 23 open source, reusable generative AI assets on 11 November 2026. AWS says customers such as DoorDash and Ricoh cut build time by using them, and two new agent starter kits are included.
If you build RAG, document processing or agent apps on AWS, you can start from reviewed code instead of a blank repo. The customer numbers are AWS's own, so test any asset against your workload before you commit.
AWS published a post on 11 November 2026 describing a catalog of 23 open source, reusable generative AI assets. The assets come from the AWS Generative AI Innovation Center, which AWS has folded into AWS Frontier AI Engineering & Services. This is AWS's own announcement.
What happened
In its Machine Learning Blog post, AWS argues that teams keep rebuilding the same parts: RAG pipelines, document processors and agent scaffolding. It names three causes: no single place to find existing work, no signal of quality, and no culture of reuse.
AWS says its assets are validated for quality and security before publication, ship with deployment guides, and are maintained over time. The group has published such assets since 2023.
Key details
- Catalog size: 23 open source assets, covering RAG components, document ingestion, agent templates, evaluation frameworks and prompt management patterns.
- DoorDash: according to an AWS case study, it built a voice-operated contact center agent on the Contact Center GenAI Agent asset. The post says it handles over 100,000 calls per day with 2.5-second response latency and a 50x increase in testing capacity.
- DoorDash timeline: AWS says using Amazon Bedrock cut development time by 50 percent, and the team was ready for live testing in 2 months.
- Lyft: it also built on the Contact Center GenAI Agent asset, for riders and drivers. The post gives no numbers for Lyft.
- Ricoh: using the Accelerated Intelligent Document Processing asset, it reports a 7-fold throughput increase, from 10,000 to 70,000 documents monthly. Engineering hours per deployment fell from 80 to under 5, and onboarding fell from 4–6 weeks to 2–3 days.
- Fullstack AgentCore Solution Template (FAST): a starter repository with a React frontend wired to Amazon Bedrock AgentCore. AWS says it deploys in under 30 minutes with infrastructure as code, and uses Amazon Cognito for authentication.
- Agentic Chatbot Accelerator: deploys with the AWS CDK and is built on Strands Agents and Amazon Bedrock AgentCore.
Why it matters
For developers, the useful part is a set of maintained starting points that are already security-reviewed, according to AWS. FAST and the chatbot accelerator target agent apps on AgentCore. The document processing and contact center assets target common enterprise jobs.
The results come from AWS and its case studies. The post does not show independent testing. Treat the figures as vendor claims and check fit, licence and maintenance status in each repository.
The post also includes a walkthrough of using an asset, which this summary does not cover.
FAQ
What is the AWS reusable asset catalog? It is a set of 23 open source generative AI building blocks from the AWS Generative AI Innovation Center. They include RAG components, document ingestion, agent templates and evaluation frameworks.
What is the Fullstack AgentCore Solution Template? FAST is a starter repository that connects a React frontend to Amazon Bedrock AgentCore. AWS says it deploys in under 30 minutes using infrastructure as code.
Are the DoorDash and Ricoh results independent? No. The figures come from AWS's post and case studies about its own assets.
Checked claims
| Claim | Status | Evidence |
|---|---|---|
| AWS published a catalog of 23 production-ready assets from its Generative AI Innovation Center. | confirmed | access a curated catalog of 23 production-ready assets from the AWS Generative AI Innovation Center |
| The Innovation Center has published open source reusable assets since 2023. | confirmed | Since 2023, the AWS Generative AI Innovation Center has published open source reusable assets. |
| According to AWS, DoorDash's contact center solution handles over 100,000 calls per day with 2.5-second response latency. | confirmed | the solution handles over 100,000 calls per day with 2.5-second response latency |
| AWS says Ricoh cut engineering hours per deployment from 80 to under 5. | confirmed | Engineering hours per deployment dropped by over 90 percent (80 hours to under 5). |
| AWS says FAST deploys in under 30 minutes using infrastructure as code. | confirmed | FAST deploys in under 30 minutes using infrastructure as code (IaC) |
Found a mistake? Use Report above; an editor reads every one. Editorial policy
Question for readersWhich part of your generative AI stack do you rebuild most often from scratch, and what would make you trust a shared asset instead?
Discussion · 0
No comments yet. Start the discussion.
Sign in to join the discussion.