Grepai Workspaces

If you're juggling multiple related codebases like a monorepo, microservices, or a frontend/backend split, this lets you group them into a single searchable workspace. You can search across all projects at once or narrow down to specific ones when you know where to look. The workspace commands are straightforward: create a workspace, add project paths, run the indexer, and search with the workspace flag. It's basically GrepAI's answer to "I need to find where this shared type is used across five different repos." The cross-project search results show which project each match comes from, which saves you from hunting through multiple index outputs.

skills.sh installs: 711
GitHub stars: 20
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add yoanbernabeu/grepai-skills --skill grepai-workspaces --agent claude-code

Installs into .claude/skills of the current project.

Files

GrepAI Workspaces

This skill covers using GrepAI workspaces to manage multiple related projects with a unified search index.

When to Use This Skill

Working with monorepos
Searching across multiple related projects
Managing microservices architecture
Organizing large codebases

What are Workspaces?

Workspaces allow you to:

Group multiple projects together
Index them into one shared vector store
Search or trace across all projects at once
Or search specific projects within the workspace

Prerequisites

Workspaces need a vector store that several projects can share:

PostgreSQL with pgvector
Qdrant

The GOB backend is not supported: it is a local file per project, so it cannot be shared.

Creating a Workspace

# Interactive prompts (backend, then embedding provider)
grepai workspace create my-workspace

# Qdrant + Ollama, all defaults
grepai workspace create my-workspace --yes

# Qdrant + OpenAI (API key read from OPENAI_API_KEY, official endpoint)
grepai workspace create my-workspace --backend qdrant --provider openai --model text-embedding-3-small

SKILL.md on GitHub