trovex vs Claude Projects
Claude Projects is uploaded knowledge inside Claude.ai. trovex serves your live repo's docs to any agent over MCP. Here's the honest comparison.
Claude Projects lets you attach documents to a project's knowledge so Claude uses them across chats — hosted inside Claude.ai, with knowledge that's whatever you upload. trovex serves your live repo's markdown to any coding agent over MCP, returning the one canonical doc per query with a freshness marker, locally, for about 60% fewer tokens per lookup. Claude Projects is a curated, manually-maintained knowledge base in one product; trovex tracks the repo as it changes and works across clients.
What is Claude Projects?
Claude Projects is a feature of Claude.ai for grouping chats around a body of knowledge. You upload documents and instructions into a project, and Claude draws on them across the conversations in that project. It's a clean way to give Claude a stable, curated context for a topic — inside Claude.ai.
Where trovex is different
trovex doesn't ask you to upload and re-upload docs. It indexes your repo's markdown where it lives, returns the single canonical doc that answers a query (with a canonical / stale / duplicate marker so the current one wins), reads at the section level, and lets agents write records back. And it isn't tied to one product — it serves any MCP client (Claude Code, Cursor, Windsurf, Cline) locally, so the same source of truth follows your agents across tools.
| Capability | Claude Projects | trovex |
|---|---|---|
| Where it lives | Inside Claude.ai (hosted) | Local, any MCP client |
| Knowledge source | Documents you upload | Your live repo's markdown |
| Stays in sync with the repo | — manual upload | ✓ re-index, no re-upload |
| Freshness signal | — | ✓ canonical / stale / duplicate |
| One canonical answer per query | ~ whole knowledge in context | ✓ section-level |
| Cross-client | — Claude only | ✓ any MCP client |
When is Claude Projects the right choice?
When you work in Claude.ai and want a curated, stable set of reference docs for a topic, Claude Projects is simple and effective. Reach for trovex when your knowledge is a living repo, when you want the current doc routed per query with a freshness signal instead of re-uploading files, or when your agents span more than one client.
FAQ
What is the difference between trovex and Claude Projects?
Claude Projects attaches uploaded documents to a project so Claude uses them across chats — hosted inside Claude.ai. trovex serves your live repo's markdown to any agent over MCP, returning the one canonical doc per query with a freshness marker, locally. Projects is a curated upload; trovex tracks the repo as it changes and works across clients.
Does Claude Projects stay in sync with my repo?
Not automatically — you upload and update the documents yourself, so the knowledge can drift, and there's no canonical/stale signal. trovex indexes the repo's markdown and marks each result canonical, stale, or duplicate, so an agent reads the current doc without manual upkeep.
Can I use trovex outside Claude?
Yes. trovex is an MCP server — it works with Claude Code, Cursor, Windsurf, Cline, and any MCP client. Claude Projects lives inside Claude.ai.
Your live repo, not re-uploaded files.
trovex is in private beta — request access and serve the current doc per query, in any client.
Open source (AGPL-3.0 core, MIT CLIs). Local-first — your docs never leave your machine.