Blog category
Production-grade guidance on Amazon Bedrock, SageMaker, Amazon Q, and multi-agent AI systems. Covers RAG pipelines, fine-tuning tradeoffs, guardrails, multi-tenant GenAI, and cost optimization for LLM workloads on AWS.
On 23 Sep 2026 AWS added Salesforce and Zendesk connectors. The managed Salesforce crawl is 4 entity flags, not the classic preview's 21 CRM objects. New customer-managed Salesforce connectors stop on 30 Sep 2026.
Managed Knowledge Base Web Crawler, GA with the service on 17 Jun 2026, follows up to 300 URLs per minute from 10 seeds or 3 sitemaps. A 10,000-URL help center is at least 34 minutes at that cap, and the crawler still has no per-page ACL.
On September 22, 2026 AWS made GPT-6 Sol and GPT-6 Luna generally available on Bedrock. OpenAI’s internal eval found Sol makes about half as many factual mistakes as GPT-5.6 Sol. Here is the lane routing.
TypeSafe priced Jev at $0.042/MTok with claimed 70–500 ms decisions on 15 Sep 2026. On AWS it is a System-1 sidecar, not a Bedrock replacement — reuse the Gateway ~180→95 ms canary and the ~$791/mo AgentCore silhouette as platform context, not as a Jev outcome.
On August 19, 2026 AWS enabled live-web fetch for Bedrock Web Search in 3 US Regions. Default external_web_access is true; AmazonBedrockFullAccess still does not grant ExternalWebAccess — and the 403 does not fail the request.
On Aug 7, 2026, AWS shipped aws-auth as a core Cognito skill in the Agent Toolkit—603 installs on skills.sh within days, 9 reference packs under skills/core-skills/aws-auth, and a hard warning that update-user-pool-client is a full replace.
On August 5, 2026 DynamoDB GA’d native vector search — single-digit ms at 99%+ recall, up to 4096 dims, SearchVectors topK ≤ 100. Decision guide vs S3 Vectors, OpenSearch, and Bedrock Knowledge Bases.
On June 15, 2026 Grok 4.3 GA on Bedrock Mantle at $1.25/$2.50 per 1M. On August 19, 2026 Grok 4.6 added Converse + US Geo/Global CRIS at $2.00–$2.20 / $6.00–$6.60 — do not jump on name alone.
On July 30, 2026 AWS cut Bedrock on-demand prices 80% for GPT-5.6 Luna and 20% for Terra — Luna is now $0.22/$1.32 per 1M tokens. Here is the routing math and what not to re-price overnight.
Quick Suite GA October 2025; Q Business and Kendra freeze new customers after July 30, 2026. A 420-seat professional-services firm paused a $38K/year Kendra quote and cut evaluation from 9 weeks to 3 by starting on Quick instead.
On a fraud-scoring team (~450 RPS peak, ml.g5.2xlarge), inference components plus April 2026 capacity-aware instance pools cut endpoint provisioning failures from 6 retries to 0 — p99 held at 68 ms through a g5→g6 fallback event.
The official aws/agent-toolkit-for-aws repo ships 43 atomic Agent Skills across 13 category folders—plus aws-core, aws-agents, and aws-data-analytics plugins. Here is why that bundle matters for IAM and audit posture, how the tree fits together, and how to pair it with the May 6, 2026 GA AWS MCP Server.