Best for
- Developers
- Technical teams
- Users needing flexible AI tool integration
- Users who want a self-hosted AI chat solution
K
Keinsaas Navigator
Community
A better open-source AI chatbot with multi-model, MCP tools, and visual workflows.
GitHub source repository ↗
32FMRS · D
This MCP server is actually a full AI chat application that acts as an MCP client (can connect to other MCP servers) and also provides its own MCP-related tools. It is ideal for users who need to integrate multiple AI capabilities and tools into one interface. Due to its feature richness, deployment and configuration are relatively complex, but it offers detailed documentation and quick deployment options.
Keinsaas Navigator (formerly Better Chatbot) is an open-source AI workspace designed to provide a feature-rich AI chat interface for individuals and teams. It integrates multiple major LLM providers (such as OpenAI, Anthropic, Google, xAI, Ollama, etc.) and supports the MCP protocol, allowing users to connect external tools and services. The platform also offers image generation, real-time voice assistant, visual workflows, custom agents, tool presets, and collaboration sharing. It is built with Next.js and Vercel AI SDK and can be easily deployed to Vercel or self-hosted via Docker.
Tools
mcp
Connect and invoke tools from external MCP servers via the MCP protocol.
web-search
Built-in web search powered by Exa AI for semantic search and content extraction.
js-executor
A simple tool for executing JavaScript code.
image-generation
Generate and edit images using AI models.
data-visualization
Create interactive tables and charts with sorting, filtering, and export options.
Setup
- Clone the repository. 2. Run
pnpm ito install dependencies (generates .env file). 3. Fill in at least one LLM provider API key (e.g., OPENAI_API_KEY) in the .env file. 4. Optionally, start a PostgreSQL database (pnpm docker:pgor use Docker Compose). 5. Runpnpm build:local && pnpm startto start the app. 6. Open http://localhost:3000 to begin.
Fit and risk
Not for
- Non-technical users, as deployment and configuration require a certain level of expertise
- Enterprises needing official support
- Users who prefer a lightweight chat interface may find it too complex
Required permissions
- Read environment variables for API keys
- Access to configured database (PostgreSQL)
- Access external APIs (LLM, Exa, etc.) as configured
- If enabled, access to file storage (Vercel Blob)
- Execute tools provided by MCP servers (e.g., browser control)
Risks and side effects
- User-configured MCP servers may execute dangerous operations; authorize with caution
- API keys stored in environment variables; ensure server security
- Code execution tools may run malicious code; restrict usage
- Data storage and privacy depend on self-hosting or cloud deployment
Troubleshooting
- Ensure all required API keys are correctly set.
- Verify that PostgreSQL is running and the connection URL is correct.
- Check the .env file for proper configuration.
- Review server logs for error messages.
- If using Docker, check that containers are running correctly.
Use cases
As a full-featured AI chatbot for individuals or teams
Integrate external tools via MCP, such as browser automation, code execution, etc.
Create visual workflows and call them as tools in chat
Use real-time voice assistant for interactive conversations
Generate and edit images with built-in image generation capabilities
Supported clients
Supported clients have not been confirmed yet.