ContextLayer
Connect your coding agent to shared team context—knowledge graph, sessions, tasks, and documents—so solo builders and small AI teams stop re-explaining the same project state every session.
Overview
ContextLayer is an MCP server for the Build phase that exposes DotNova’s knowledge graph, sessions, tasks, and documents to your agent over streamable HTTP.
What is this MCP server?
- Streamable HTTP remote at https://api.dotnova.io/mcp (schema version 0.1.0)
- Bearer JWT from app.dotnova.io Settings → API Keys
- Knowledge graph, sessions, tasks, and documents in one context layer
- Hosted app at https://app.dotnova.io with GitHub repo DotNova/context-layer
- Server schema version 0.1.0
- Remote MCP URL: https://api.dotnova.io/mcp
- Auth: required Bearer JWT header
What problem does it solve?
Agents lose thread across sessions and teammates because project knowledge lives in chats, tickets, and docs with no shared graph the model can query.
Who is it for?
Solo builders or tiny AI teams already on DotNova who want centralized context wired into Claude Code or Cursor via official MCP remotes.
Skip if: Builders who want a local-only, no-account context store or who do not want a hosted Bearer-token dependency on api.dotnova.io.
What do I get? / Deliverables
After you register the remote MCP URL and JWT, your agent can pull structured team context instead of you pasting the same background every time.
- Agent-callable access to knowledge graph, sessions, tasks, and documents
- Authenticated remote MCP connection to api.dotnova.io
Recommended MCP Servers
Journey fit
Context infrastructure is installed while you wire agents and workflows during product build, even though the same context helps later iteration. Canonical shelf is agent-tooling because the server exposes remote MCP access to DotNova’s context layer, not a one-off app feature.
How it compares
Hosted context MCP integration, not an agent skill or static SKILL.md workflow.
Common Questions / FAQ
Who is ContextLayer for?
Indie builders and small AI teams using DotNova who need shared knowledge, sessions, tasks, and documents available to coding agents through MCP.
When should I use ContextLayer?
When you are integrating agents during build (or iterating in production) and repeated context dumps are slowing you down.
How do I add ContextLayer to my agent?
Sign in at https://app.dotnova.io, copy an API token from Settings → API Keys, then add the remote MCP server https://api.dotnova.io/mcp with Authorization: Bearer <token> in your Claude Code or Cursor MCP config.