About Convika
Convika is a hosted service for creating, publishing, and tracking AI-made landing pages, aimed at people managing pages and signups. It provides private previews before publishing, version rollback, forms, visitor analytics, and custom-domain setup.
Key features
- 01Create and edit landing pages
- 02Preview before publishing
- 03Publish, unpublish, or roll back
- 04Collect and export form submissions
- 05Review visitor and campaign analytics
- 06Set up custom domains
Use cases
- 01Preview an AI-written page before launch
- 02Publish a campaign signup page
- 03Review traffic sources and signups
- 04Host a page on a custom domain
MCP support
The hosted MCP server uses Streamable HTTP and OAuth 2.1. It exposes tools for managing landing pages, previews and publishing, forms, analytics, and domains.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.convika.com/mcpclaude mcp add --transport http com-convika-convika https://mcp.convika.com/mcp{"mcpServers":{"com-convika-convika":{"type":"http","url":"https://mcp.convika.com/mcp"}}}Open in Cursor{"servers":{"com-convika-convika":{"type":"http","url":"https://mcp.convika.com/mcp"}}}Open in VS Code{"mcpServers":{"com-convika-convika":{"type":"http","url":"https://mcp.convika.com/mcp"}}}{"mcpServers":{"com-convika-convika":{"serverUrl":"https://mcp.convika.com/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEConvika MCP Server
Publish AI-made landing pages on your own domain, with a signup form and visitor stats.
mcp-name: com.convika/convika
Convika hosts landing pages made with AI. You can upload the HTML file ChatGPT, Claude, or Gemini wrote, generate a page in the dashboard, or connect your AI client here and let it save and publish the page for you. A page can carry a form that keeps every entry, and every page gets visitor stats: visits, traffic sources, UTM campaigns, button clicks, and signups.
This repository holds the public documentation and the server manifest for the hosted Convika MCP server. The service itself is closed source.
- MCP endpoint:
https://mcp.convika.com/mcp(Streamable HTTP, OAuth 2.1) - Official MCP Registry:
com.convika/convika - Docs: https://convika.com/docs/
- Sign up (free, no card): https://app.convika.com/signup
Quick start
Claude (web and desktop)
- Settings → Connectors → Add custom connector
- Enter
https://mcp.convika.com/mcp - Sign in to your Convika account, or create one for free
Claude Code
Install the Convika plugin. It adds this server and a publish-page skill, which turns an HTML file in your repo or a page Claude wrote into one document, shows you a private preview, and asks before it publishes.
/plugin marketplace add tomoyaf/convika-mcp
/plugin install convika@convika
Then run /mcp, select the Convika server, and choose Authenticate. Details: plugins/convika
To add the server without the skill:
claude mcp add --transport http convika https://mcp.convika.com/mcp
Codex, Cursor, and other MCP clients
{
"mcpServers": {
"convika": {
"url": "https://mcp.convika.com/mcp"
}
}
}
ChatGPT can connect in developer mode with the same URL.
Sign-in uses OAuth 2.1 with dynamic client registration and PKCE. Your client opens a browser window to sign in, so there are no API keys to manage.
What your AI can do
| Area | Tools |
|---|---|
| Create and edit pages | create_landing_page_preview, create_landing_page, create_site, list_sites, get_site, update_site, save_page_content, get_page_structure |
| Preview and publish | create_preview, publish_site, rollback_site, unpublish_site, create_claim_code |
| Forms and goals | create_form, get_form_submissions, export_form_submissions, set_goals, list_goals |
| Visitor stats | get_analytics_summary, get_analytics_detail, get_portfolio_analytics, analyze_site_performance, analyze_source_quality, query_timeseries, query_segments, query_cta_performance, detect_trend |
| Domains and URLs | set_site_url, setup_custom_domain, get_domain_status, list_domain_routes, disable_domain_route, start_domain_control_verification, check_domain_control_verification, activate_domain_handover |
| Other | ping, get_referral_link |
The usual flow: your AI writes the page and calls create_landing_page_preview, which returns a private preview link. Nothing goes live until you ask it to publish. Each save is a new version, so a rollback is one call. Published pages keep serving even if the dashboard is down.
Example prompts
"Make a landing page for my coffee shop's opening with a signup form, and show me the preview."
"Publish it."
"How did the page do this week? Which source brought the most signups?"
"Put this page on www.mycoffeeshop.com."
Plans
- Free: $0, no credit card. 10 pages, 10,000 pageviews a month, 50 form entries a month, 1 custom domain (a subdomain such as
www), 90 days of visitor stats, and a small "Powered by Convika" badge. - Pro: $12 a month, or $108 a year. 50 pages, unlimited form entries and custom domains, no badge, and a 14-day free trial.
Details: https://convika.com/pricing/
Support
- Problems with the MCP server or these docs: open an issue in this repository.
- Product support: https://convika.com/support/
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.