Comparison

DearTech-OS vs Claude Projects

Claude Projects are file attachments and custom instructions inside Claude's chat surface. DearTech-OS is a typed knowledge graph your AI can search and traverse, accessible from any MCP-compatible tool. Both let you give AI more context. The difference is what shape that context takes and where it can flow.

Last updated:

Side by side

Where they differ

CriterionClaude ProjectsDearTech-OS
AI tool supportClaude onlyClaude, ChatGPT, Cursor, Codex, any MCP-compatible tool
Knowledge structureAttached files + custom instructions, flatTyped graph: concepts, patterns, decisions, status, confidence, relationships
Graph search & traversalNo, files are chat-context onlyYes. AI can search nodes, traverse relationships, audit sources
Vendor portabilityAnthropic-hosted, accessed through AnthropicMarkdown files canonical on your disk. Leave any vendor without losing knowledge
Source attributionFiles attached, no per-claim sourcingEvery claim carries source, status, confidence, ownership
Beyond chatStays inside Claude conversationsFeeds dashboards, board reports, agents, other AI tools
Query interfaceChat UI in Claude onlyPlain English in any AI tool + programmatic MCP queries + CLI search
Sharing model (Pro)Per-userShared by design
Sharing model (Team / Enterprise)Shared within Claude workspaceShared across all AI tools, not just one vendor
Best forTeams standardized on Claude for chat-based AI workflowsFounder-operator teams using multiple AI tools who need graph-queryable shared context

Where Claude Projects excels

  • Native to Claude. Zero setup if your team already uses Claude.
  • Strong for chat-based AI workflows
  • Available across Pro, Team, and Enterprise plans (with sharing on paid tiers)
  • File upload plus custom instructions cover most chat use cases
  • Anthropic-managed infrastructure

Where DearTech-OS excels

  • Works with any AI tool via MCP: Claude, ChatGPT, Cursor, Codex
  • Typed knowledge graph that AI can search and traverse, not just read
  • Markdown files canonical on your disk. Leave any vendor without losing your knowledge layer.
  • Source-grounded answers (every claim carries source, status, confidence)
  • Knowledge feeds beyond chat: dashboards, reports, agents, other tools
  • Programmatic query interface for building on top

Decision

When to choose which

Choose Claude Projects when

  • Your team uses Claude (and only Claude) for AI work
  • File attachments and custom instructions cover your needs
  • Chat is your only AI surface, no agents, dashboards, or pipelines querying knowledge
  • You don't need a typed graph your AI can traverse and search

Choose DearTech-OS when

  • You use multiple AI tools (Claude, ChatGPT, Cursor, etc.)
  • You want a typed graph your AI can search and traverse, not just attached files
  • You want knowledge portable across vendors and surviving any vendor change
  • You want company context to feed dashboards, reports, and agents, not just chat
  • Non-technical teammates query through their AI tool of choice

Use them together

DearTech-OS is AI-tool-agnostic

DearTech-OS doesn't replace your AI tool. It gives that AI tool more to work with. If your team already uses Claude Projects for chat-native workflows, keep using them. DearTech-OS sits underneath as the typed knowledge layer your Claude (and ChatGPT, and Cursor) can search and traverse, with shared source-grounded answers across every AI surface in your stack.

FAQ

Common questions

Can I use both Claude Projects and DearTech-OS?

Yes. Many teams use Claude Projects for chat-native workflows and DearTech-OS as the structured, graph-queryable layer that works across all their AI tools.

Does DearTech-OS work with Claude?

Yes. DearTech-OS exposes the company knowledge graph through an MCP server, which Claude (Pro, Team, or Enterprise) can query natively. It also works with ChatGPT, Cursor, Codex, and any other MCP-compatible AI tool.

We are on Claude Team or Enterprise. Projects are already shared. Why do we need DearTech-OS?

Even shared Claude Projects work only inside Claude's chat surface. DearTech-OS gives you a typed knowledge graph your AI can search and traverse, exposes that graph to any MCP-compatible tool, and keeps your knowledge in markdown files on your disk, so it can flow into dashboards, agents, and other AI tools, and survive any vendor change.

What does 'graph search and traversal' actually mean for a non-technical teammate?

When a non-technical teammate asks 'show me everything we know about ICP X,' DearTech-OS doesn't just return a list of attached documents. It traverses the typed graph, pulling related concepts (objections, customer evidence, positioning, sales learnings) with their sources and status. That is something flat file attachments cannot do.

Does the knowledge in DearTech-OS lock me to one vendor?

No. The knowledge layer is markdown files on your disk. The graph is parsed from those files. The MCP server exposes the graph to whichever AI tools you choose. If you stop using DearTech-OS or switch AI vendors, your knowledge stays exactly where it is, fully portable.

Want a knowledge graph your AI can actually traverse?

In a 30-minute call, we'll map where your company knowledge is scattered today and identify the first DearTech-OS layer worth building.

Let's chat