Knowledge Graph

Wiki-links, embeddings, semantic search, and RAG for connecting everything in your workspace.

Trellis uses [[wiki-links]] to create bidirectional references across all workspace entities:

  • [[TRL-5]] — Link to an issue
  • [[src/engine.ts]] — Link to a file
  • [[src/engine.ts#createIssue]] — Link to a specific symbol
  • [[decision:DEC-1]] — Link to a decision trace
import { parseFileRefs, resolveRef, RefIndex } from "trellis/links";

const refs = parseFileRefs("src/engine.ts", source);
const resolved = resolveRef(ref, context);

const index = new RefIndex();
index.indexFile("README.md", refs, context);
index.getIncoming("issue:TRL-5"); // All refs pointing to this issue

Trellis can embed issues, milestones, files, code entities, and decisions into a vector store for intelligent discovery:

trellis ask "authentication code"
trellis ask "show me auth code" --rag
import { EmbeddingManager } from "trellis/ai";

const manager = new EmbeddingManager(engine, { dbPath: "embeddings.db" });
await manager.reindex();
const results = await manager.search("auth flow");

RAG Context

Generate context for LLMs from your codebase:

const context = await manager.getRAGContext("authentication implementation");
// Returns relevant code, issues, and decisions as structured context

Cross-Repository References

Link knowledge across repositories:

index.addCrossRepoLink("frontend", "proj:app", "backend", "api:users");