Install and configure RAG primitives in a TypeScript codebase.
npx skills add https://github.com/BetterStacks/unrag --skill unrag
Please help me install this Agent Skill. Skill: unrag Source: https://github.com/BetterStacks/unrag/tree/main/packages/unrag/skills/unrag Command: npx skills add https://github.com/BetterStacks/unrag --skill unrag
This Skill simplifies the integration of Retrieval-Augmented Generation (RAG) into your TypeScript projects by providing auditable, drop-in source code primitives instead of complex frameworks.
Initialize unrag in your project by running bunx unrag@latest init.
High-intent search queries and answers about installing and using this skill.
You can add retrieval-augmented generation by vendoring auditable source files directly into your TypeScript project, using simple primitives like `ingest()` and `retrieve()` for local-first development without external frameworks.
Document ingestion and embedding generation use swappable interfaces for embedding providers and extractors, processing documents and storing vectors locally within your repository through simple function calls.
Yes, embedding providers and vector store adapters are swappable. The implementation provides simple interfaces for these components, enabling you to customize your RAG pipeline's storage and embedding behavior within your codebase.
The best way to integrate a search and retrieval system into a documentation site is by vendoring RAG primitives directly into your codebase, ensuring minimal setup, full code ownership, and no dependency on external services.
You initialize and configure RAG capabilities by running `bunx unrag@latest init` in your TypeScript project, which sets up vendored source files for document ingestion, embedding generation, vector storage, retrieval, and reranking.
No external services or dependencies are required. The implementation operates local-first, keeping your vector database and RAG implementation entirely within your repository for full code ownership and auditability.
MINT-IACoordinate parallel agents to generate seven codebase documents in .planning/codebase/.
opencodosLocate past conversations and decisions in an Obsidian Vault using hybrid search.
constructive-ioCreate and manage pgvector databases in PostgreSQL with Ollama for RAG pipelines.
cloudlinessHost a scalable managed vector database for similarity search.
