Qelvra, inside the AI you already use.
Our remote MCP server gives Claude, ChatGPT and Cursor direct access to your sites, your action plan and the Qelvra agent — with the same live data and playbooks as the app.
Endpoint
https://qelvra.in/api/mcp
Transport
Streamable HTTP
Auth
Authorization: Bearer qv_…
Two minutes, four clients.
First, create an API key in Settings → API keys inside the app. Keys start with qv_ and are shown once — keep them somewhere safe.
Claude Code
Run once. Qelvra's tools are then available in every Claude Code session.
claude mcp add --transport http qelvra https://qelvra.in/api/mcp \
--header "Authorization: Bearer qv_your_api_key"Claude Desktop
Settings → Developer → Edit config. Or add it as a custom connector with the same URL and header.
{
"mcpServers": {
"Qelvra": {
"url": "https://qelvra.in/api/mcp",
"headers": {
"Authorization": "Bearer qv_your_api_key"
}
}
}
}Cursor
Add to your project's .cursor/mcp.json (or the global one), then enable Qelvra under Settings → MCP.
{
"mcpServers": {
"Qelvra": {
"url": "https://qelvra.in/api/mcp",
"headers": {
"Authorization": "Bearer qv_your_api_key"
}
}
}
}ChatGPT
Requires a ChatGPT plan with custom connectors (developer mode). Create a new connector with these values.
Name: Qelvra
MCP server URL: https://qelvra.in/api/mcp
Authentication: API key / Bearer
Token: qv_your_api_keyEleven tools. One that does the heavy lifting.
Most of the time your assistant only needs send_message — it hands your question to the Qelvra agent, which runs its own research and returns the answer. The rest let it read and manage your workspace.
send_messageMain entrypointAsk the Qelvra agent anything. It picks the playbook, runs the tools and replies. Start here.
get_creditsYour remaining credits and plan.
list_sitesSites connected to your account.
get_siteDetails and connection status for one site.
create_siteAdd a new site by domain.
list_actionsThe prioritised action plan for a site.
update_actionMark an action done, snoozed or dismissed.
list_conversationsYour past conversations with the agent.
get_conversation_messagesFull message history for a conversation.
list_artifactsReports, briefs and articles the agent has produced.
get_artifactFetch one artifact's full content.
Credits, limits and security
MCP access is included on Pro, Ranker and Agency. Each send_message call uses one credit, same as in the app; reading your data is free. Keys are scoped to your account, can be revoked any time, and never grant access to your Google password.