Docs agents that answer for the version the customer runs

1. 01 · Ingest

   Scope what you ingest

   Every document lands with a groupName: the sets it belongs to. Those sets are what retrieval intersects later, so the structure you choose here is the precision you get there.

   context.add

   import AlchemystAI from "@alchemystai/sdk";const client = new AlchemystAI(); // reads ALCHEMYST_AI_API_KEYawait client.v1.context.add({  context_type: "resource",  scope: "external",  source: "docs",  documents: [{    content: "API v3: tokens are issued by POST /v3/auth/token. The /v2/login endpoint is removed.",  }],  metadata: {    fileName: "v3/auth/tokens.md",    groupName: ["docs", "api", "v3"],   // the sets this belongs to  },});

2. 02 · Write

   Mark the version, not just the page

   When v3 ships, v2 pages do not disappear. Keep them in their own version scope, so v2 users still get v2 answers and v3 users never see them.

   context.add

   await client.v1.context.add({  context_type: "resource",  scope: "external",  source: "docs",  documents: [{ content: v2LoginPage, status: "deprecated" }],  metadata: {    fileName: "v2/auth/login.md",    groupName: ["docs", "api", "v2"],    lastModified: "2025-11-03T00:00:00Z",  },});

3. 03 · Search

   Search before answering

   Search intersects the scopes, subtracts superseded and duplicate content, and ranks what survives. Only that reaches the model, and the whole decision is recorded as a Context Trace.

   context.search

   const { contexts } = await client.v1.context.search({  query: "How do I get an access token?",  scope: "external",  similarity_threshold: 0.8,  minimum_similarity_threshold: 0.5,  metadata: { groupName: ["docs", "api", "v2"] },   // ∩ narrow scope});// − superseded, deduplicated → ranked → into the window// Every search is recorded as a Context Trace.const reply = await llm.respond(message, { context: contexts });

npm install @alchemystai/sdk or pip install alchemystai, both ship the same client. Full reference in the docs.