How Postman runs Agent Mode for 40 million developers on Amazon Bedrock

Postman demonstrates how to run an AI agent at scale on Amazon Bedrock by adopting a set of architectural patterns that tame tool sprawl, expose schema‑based reads, and treat context size as the primary bottleneck. The approach allows a single agent to serve 40 million developers by limiting the number of integrated tools, using OpenAPI‑style schemas to fetch data, and carefully managing Bedrock’s context window to keep latency low. The result is a production‑ready agent that can be deployed at petabyte‑scale workloads without sacrificing responsiveness.

⚡ Key Takeaways

• The agent limits tool integration to a curated set, preventing uncontrolled “tool sprawl” that would otherwise bloat request size and latency.
• Schema‑based reads are exposed via OpenAPI specifications, allowing the agent to fetch structured data directly from services with minimal parsing overhead.
• Context size is treated as the real bottleneck; the design explicitly caps prompt length to fit within Bedrock’s maximum context window, ensuring consistent performance.
• Integration is achieved by configuring Postman’s Agent Mode to point to Bedrock endpoints, passing the schema and tool list as part of the request payload.
• WhyItMatters: For engineers shipping AI agents to millions of users, this pattern shows how to keep tool integration manageable, reduce latency by respecting context limits, and leverage Bedrock’s managed service to scale without custom infrastructure.
• TechnicalLevel: Advanced
• TargetAudience: ML Engineers, RAG Practitioners
• PracticalSteps:
• Configure

For engineers shipping AI agents to millions of users, this pattern shows how to keep tool integration manageable, reduce latency by respecting context limits, and leverage Bedrock’s managed service to scale without custom infrastructure.

✅ Practical Steps

1. Configure

Want the full story? Read the original article.