Developer Changelog Retrieval: Freshness, Tenant Isolation, Citation Control
Short answer: use vector retrieval for durable changelog context, web retrieval for freshly published changes, and put both behind a tenant-scoped contract so you can switch providers without rewriting the tracker. That decision is less about picking a fashionable search engine and more about deciding what an answer is allowed to cite. A developer changelog tracker has two clocks: the indexed…
To create tenant-safe retrieval for a developer changelog tracker, prioritize the user-facing answer first. Begin by returning evidence objects instead of just strings for query results. Each object must include a stable document ID, tenant ID, source URL, publication timestamp, text span, and retrieval score. The answer layer should reject any object that doesn't match the requested tenant ID or falls outside the specified timestamp window.
Regardless of the implementation method—whether using an in-house index, hosted vector database, or web search API—the contract should unify the retrieval operations. In TypeScript, this can be achieved with clear interfaces: RetrievalRequest for request parameters, Evidence for the returned objects, and Retriever for the retrieval function.
Implement a function to list Infrai collections using a provided API key, handling rate limiting with retries. Create an async function to retrieve evidence for a specific tenant by filtering the provider's results based on tenant ID and limiting the number of results. Log tenant ID, query ID, candidate count, and evidence IDs, but avoid logging raw protected text by default to ensure tenant safety.
Implement tenant ID validation at the start to fail closed if it's missing before calling the adapter.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.