Alcove FAQs

Question 1

What problem does Alcove solve for AI coding agents?

Accepted Answer

Alcove addresses the challenge of AI agents lacking project-specific context and suffering from context bloat. It provides on-demand access to your private project documentation, ensuring agents retrieve only the most relevant information without being overloaded.

Question 2

How does Alcove prevent context bloat when working with AI agents?

Accepted Answer

Instead of injecting all documentation into the AI agent's context window, Alcove employs BM25 ranked search. Agents can query for specific information, and Alcove retrieves only the most relevant documents, significantly reducing context size and improving efficiency.

Question 3

Is my private project documentation secure with Alcove?

Accepted Answer

Yes, security is a core feature. Alcove ensures that your private documents remain on your local machine, never exposed in public repositories or externally. It operates entirely locally to maintain your data's privacy and confidentiality.

Question 4

Can I use Alcove with multiple AI agents and across different projects?

Accepted Answer

Absolutely. Alcove is designed for multi-agent and multi-project environments. It offers a standardized documentation layer, allowing one setup to serve private docs to any MCP-compatible AI agent across various projects with automatic current working directory (CWD) detection.

Question 5

How does Alcove differ from simply putting documentation in files like `CLAUDE.md`?

Accepted Answer

While `CLAUDE.md` is suitable for short conventions or agent behavior rules, Alcove is built for scalable project knowledge like architecture, decisions, and runbooks. It avoids context bloat by allowing agents to search for specific needs, unlike `CLAUDE.md` which injects all content on every run, leading to slower and more expensive interactions.

Alcove

Alcove

Key Features

Use Cases

Key Features

Use Cases