IDE Integrations
Cognova uses MCP (Model Context Protocol) as its single integration layer. Every IDE and AI agent that supports MCP can connect to your workspace — no SDKs, no REST APIs, just one protocol endpoint.
How It Works
Your workspace exposes an MCP server at https://cognova.dev/mcp. Authenticate with an API key, and your client automatically discovers all available tools, resources, and capabilities.
Choose Your Setup
Plugin — Claude Code (recommended)
The Cognova plugin includes automated hooks that make your workspace proactive:
- Memory recall — silently searches memories for project context on session start
- Auto-remember — stores design decisions, patterns, and gotchas at end of turn
- Duplicate guard — checks for existing tasks before creating new ones
/plugin marketplace add cognova-dev/cognova-skill
/plugin install cognova@cognova-dev
Then run /cognova to configure your API key. Full details: Claude Code Integration
Cross-Agent Skill — skills.sh
Works in Claude Code, Cursor, Codex CLI, Gemini CLI, GitHub Copilot, and 40+ other agents. Installs the skill (instructions for the agent) but not hooks.
npx skills add cognova-dev/cognova-skill
MCP Direct — Any Client
For Cursor, Windsurf, or any MCP-compatible client, configure the endpoint directly:
{
"mcpServers": {
"cognova": {
"url": "https://cognova.dev/mcp",
"headers": {
"Authorization": "Bearer cg_YOUR_KEY"
}
}
}
}
IDE-specific guides: Cursor Setup | Full MCP Reference
What You Get
Tools (19)
| Category | Read (Free+) | Write (Starter+) |
|---|---|---|
| Tasks | list_tasks, get_task | create_task, update_task, complete_task |
| Projects | list_projects | create_project, update_project |
| Agents | list_agents | ask_agent (uses credits) |
| Memory | recall | remember |
| Knowledge | list_knowledge, get_knowledge | — |
| Secrets | list_secrets | get_secret, set_secret |
| Billing | get_balance | — |
| Docs | search_docs | — |
All tools are prefixed with cognova_ (e.g., cognova_list_tasks).
Resources (5)
| URI | Description |
|---|---|
cognova://tasks/active | All non-completed tasks |
cognova://knowledge/{path} | Knowledge file by virtual path |
cognova://agents/{agentId} | Agent details with tools |
cognova://docs | Documentation table of contents |
cognova://docs/{slug} | Individual doc page |
Access Levels
| Level | Plans | Capabilities |
|---|---|---|
| Read | All plans (Free+) | List and search across all data |
| Write + Agent | Starter+ ($19/mo) | Create, update, delete, and invoke agents |
Next Steps
- Create an API key if you haven't already
- Choose your setup above and connect
- Start using
cognova_*tools in your IDE