https://aivax.net/blog/ Field notes on AI infrastructure, retrieval and model operations. en https://aivax.net/blog/research-is-a-pipeline-search-discovers-fetch-reads/ https://aivax.net/blog/research-is-a-pipeline-search-discovers-fetch-reads/ Sat, 12 Sep 2026 00:00:00 GMT A server-side research workflow should separate source discovery from source reading, preserve the URLs and extracted content as evidence, and evaluate each stage independently. https://aivax.net/blog/upgrade-your-harness-with-aivax-mcps/ https://aivax.net/blog/upgrade-your-harness-with-aivax-mcps/ Wed, 09 Sep 2026 00:00:00 GMT Coding agents lose context between sessions and guess at the world outside the repo. AIVAX's Collections and Web Utilities MCP servers give them writable semantic memory plus fetch and search — with scoped credentials, bounded retrieval, and explicit write control. https://aivax.net/blog/persistent-memory-is-a-write-path/ https://aivax.net/blog/persistent-memory-is-a-write-path/ Mon, 07 Sep 2026 00:00:00 GMT An agent that writes its own memory gains a second input channel that outlives the session. Recent poisoning research shows why that channel needs ownership, expiry, and review — and how AIVAX scopes it. https://aivax.net/blog/batch-is-an-admission-control-problem-not-a-queue/ https://aivax.net/blog/batch-is-an-admission-control-problem-not-a-queue/ Sun, 06 Sep 2026 00:00:00 GMT Bulk AI work fails at the boundary between the queue and the provider: rate limits, balance, validation, and overload. AIVAX Batch answers with bounded admission, per-item validation, and failure-shaped retries. https://aivax.net/blog/judge-trajectory-production-signal/ https://aivax.net/blog/judge-trajectory-production-signal/ Fri, 04 Sep 2026 00:00:00 GMT Agentic Tests already scores every turn of a simulated conversation. Applied to real production traffic, the same trajectory signal — score, at-risk state, persistent loss — tells you when a live conversation is drifting before the user gives up. https://aivax.net/blog/mcp-stateless-core-moves-session-state-into-your-application/ https://aivax.net/blog/mcp-stateless-core-moves-session-state-into-your-application/ Wed, 02 Sep 2026 00:00:00 GMT MCP 2026-07-28 removes protocol sessions and transport replay, moving durable state, retries, long-running work, and compatibility into explicit application contracts. https://aivax.net/blog/a-vector-database-is-not-a-rag-system/ https://aivax.net/blog/a-vector-database-is-not-a-rag-system/ Mon, 31 Aug 2026 00:00:00 GMT A vector database stores and searches vectorized records; a managed RAG system coordinates the wider path from source content to maintained, retrievable context. https://aivax.net/blog/mcp-is-a-trust-boundary-not-just-a-tool-catalog/ https://aivax.net/blog/mcp-is-a-trust-boundary-not-just-a-tool-catalog/ Fri, 28 Aug 2026 00:00:00 GMT Connecting an MCP server exposes executable capabilities to a model. Discovery policy, call-time authorization, result handling, and audit decide whether the connection is safe — not the tool list. https://aivax.net/blog/migrating-from-openai-assistants-without-rebuilding-the-same-coupling/ https://aivax.net/blog/migrating-from-openai-assistants-without-rebuilding-the-same-coupling/ Thu, 27 Aug 2026 00:00:00 GMT The Assistants API shutdown is an architecture migration, not an endpoint rename. Map configuration, state, execution, tools and observability before choosing Responses or an application-owned gateway design. https://aivax.net/blog/reasoning-is-a-protocol-not-just-a-model-setting/ https://aivax.net/blog/reasoning-is-a-protocol-not-just-a-model-setting/ Tue, 25 Aug 2026 00:00:00 GMT Reasoning APIs now carry summaries, signatures, encrypted state and provider-specific stream events. A production gateway must preserve that protocol, not flatten it into text. https://aivax.net/blog/aivax-gateway-moderation/ https://aivax.net/blog/aivax-gateway-moderation/ Mon, 24 Aug 2026 00:00:00 GMT AIVAX Gateway Moderation centralizes safety policy, normally evaluating incoming content before model generation while preserving a configured fallback when that separate step is unavailable. https://aivax.net/blog/introducing-agentic-tests/ https://aivax.net/blog/introducing-agentic-tests/ Sun, 16 Aug 2026 00:00:00 GMT Agentic Tests evaluates complete, goal-oriented conversations through your AI Gateway, with a simulated user and an independent judge tracking the outcome across turns. https://aivax.net/blog/semantic-search-vs-reranking/ https://aivax.net/blog/semantic-search-vs-reranking/ Wed, 29 Jul 2026 00:00:00 GMT A practical guide to diagnosing retrieval, rank fusion, and reranking—so you improve the stage that is actually failing. https://aivax.net/blog/reflex-retrieval-built-for-recurring-documents/ https://aivax.net/blog/reflex-retrieval-built-for-recurring-documents/ Wed, 22 Jul 2026 00:00:00 GMT Reflex combines semantic relevance, lexical evidence and account-scoped cache reuse for recurring-document retrieval.