live wire
▸AI · EAGLE-4 0.1.4 fixes multi-GPU vLLM collective ordering during startupEAGLE-4▸SECURITY: Apache Camel 4.23.0, 4.22.2 and 4.18.5 fix eight component vulnerabilitiesApache Camel▸AI · Red Hat shows DRA doubles llm-d KV-cache throughput by aligning GPUs and RDMA NICsRed Hat Developer▸Red Hat Developer Hub 1.9.10 ships as a Critical security updateRed Hat errata▸OpenShift Dev Spaces 3.31 adds live editor switching and DevWorkspace Operator 1.0Red Hat documentation▸OpenShift Pipelines 1.24.1 tightens Manual Approval Gate RBAC and approval integrityRed Hat release notes▸AI · Red Hat survey: 97% of UK CIOs trust agent security, but 43% can trace data flowsRed Hat▸JAVA · IBM and Kaggle launch a living benchmark for agentic Java framework migrationsIBM Research▸AI · Red Hat tests policy-derived data controls for AI agentsRed Hat▸Apache Camel pattern turns routes into bounded LLM tools and MCP servicesRed Hat Developer▸Atos tunes small models on OpenShift AI for structured German-language analysisRed Hat Developer▸Red Hat publishes checklist for evaluating third-party OpenShift applicationsRed Hat Developer▸SECURITY · OpenShift 4.21.40 patches OpenSSL, OpenSSH and PCRE2 flawsRed Hat Product Security▸SECURITY · Apache Camel fixes A2A webhook SSRF gap in 4.22.2 and 4.23.0 (Oct. 23)Apache Camel▸AI · EAGLE-4 0.1.4 fixes multi-GPU vLLM collective ordering during startupEAGLE-4▸SECURITY: Apache Camel 4.23.0, 4.22.2 and 4.18.5 fix eight component vulnerabilitiesApache Camel▸AI · Red Hat shows DRA doubles llm-d KV-cache throughput by aligning GPUs and RDMA NICsRed Hat Developer▸Red Hat Developer Hub 1.9.10 ships as a Critical security updateRed Hat errata▸OpenShift Dev Spaces 3.31 adds live editor switching and DevWorkspace Operator 1.0Red Hat documentation▸OpenShift Pipelines 1.24.1 tightens Manual Approval Gate RBAC and approval integrityRed Hat release notes▸AI · Red Hat survey: 97% of UK CIOs trust agent security, but 43% can trace data flowsRed Hat▸JAVA · IBM and Kaggle launch a living benchmark for agentic Java framework migrationsIBM Research▸AI · Red Hat tests policy-derived data controls for AI agentsRed Hat▸Apache Camel pattern turns routes into bounded LLM tools and MCP servicesRed Hat Developer▸Atos tunes small models on OpenShift AI for structured German-language analysisRed Hat Developer▸Red Hat publishes checklist for evaluating third-party OpenShift applicationsRed Hat Developer▸SECURITY · OpenShift 4.21.40 patches OpenSSL, OpenSSH and PCRE2 flawsRed Hat Product Security▸SECURITY · Apache Camel fixes A2A webhook SSRF gap in 4.22.2 and 4.23.0 (Oct. 23)Apache Camel
upstreambeat.ai
releaseAI

Open GenAI Stack 1.5 makes API selection explicit and moves MCP clients to 2.x

The upgrade broadens provider support, but integrators should audit `apis:` configuration and test custom MCP adapters before rollout.

API config changed: omitted key, empty list, and route differences.
AI-generated illustration
By The News Desk· Oct 9, 2026the quick take — two AI hosts go live when you do

Open GenAI Stack 1.5 is not a drop-in upgrade for every deployment. The Oct. 9 release collects 108 commits since 1.4.0, adds MCP 2.x client support and new inference, search and storage integrations, but also changes how the server decides which HTTP APIs to expose.

API configuration now means what it says

The sharpest change is the treatment of apis:. In 1.4, apis: [] was indistinguishable from omitting the key and could expose every API supplied by configured providers. In 1.5, an explicit empty list is authoritative: it serves no provider-backed APIs. Omitting the key still means “serve everything,” while a bare apis: value is treated like an omitted key and produces a startup warning.

Route registration also follows a new three-tier rule. The merged routing change keeps admin, inspect and providers available for stack administration; serves conversations and prompts when they are listed or when responses is enabled; and serves other APIs only when they are listed. A deployment with an explicit list that omits responses must therefore add conversations or prompts if it expects those endpoints. Configurations that include responses, and configurations that omit apis:, retain the previous route surface.

That distinction matters most for gateways and split-service topologies. The project’s test case shows that removing responses from the list also turns off /v1/responses, /v1/conversations and /v1/prompts, while unrelated configured routes remain available. The same change notes that the bundled distributions already include responses, so their HTTP surface did not require configuration edits.

MCP 2.x is supported, with limits

The MCP integration now requires the MCP 2.x client SDK, resolving to 2.2.0 in the release work. OGX migrated its client code to renamed SDK attributes, switched transport exception handling to the SDK’s httpx2 dependency, moved custom headers onto a caller-owned asynchronous client and explicitly maps an HTTP 401 during connection to its authentication-required error.

The project reports end-to-end checks for both SSE and streamable HTTP transports, including server information, tool discovery, invocation, sessions, custom headers and the 401 path. Existing /sse URLs and distribution configurations remain unchanged. The same change explicitly leaves multi-round-trip elicitation, structuredContent, cache TTLs, reserved MCP headers and W3C trace propagation for follow-up work, so “MCP 2.x support” does not yet mean every new SDK capability is implemented.

Upgrade checklist

Before moving from 1.4, integrators should distinguish an omitted apis: key from an empty list, verify every route their gateway expects, and add conversations or prompts explicitly when responses is not exposed. Teams with custom MCP adapters should also test both transports, authentication failures and any code that depends on the old McpError name or httpx exception types. The 1.5 release notes describe 45 additional fixes and new providers, but the configuration and client-SDK transitions are the changes most likely to surface during rollout.

Filed by The News Desk. Corrections: desk@upstreambeat.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.