The kitchen-sink chatbot builder for your own data

What it does Dialoqbase is a self-hostable app for building custom chatbots backed by your own knowledge base. It ingests websites, PDFs, Word documents, GitHub repositories, and even audio or video files, stores them in PostgreSQL with vector search, and generates context-aware responses using the language model you configure. You can talk to the bot through a web interface, an embeddable script, or beta-grade Telegram, Discord, and WhatsApp integrations.

The interesting bit The ingestion pipeline is the real workhorse. Most open-source knowledge-base bots top out at text and PDFs, but Dialoqbase will also slurp in GitHub repositories, YouTube links, MP3s, and MP4s—effectively turning your digital hoard into something queryable.

Key highlights

• Broad LLM support: OpenAI, Anthropic, Google, Fireworks, Local AI, and Ollama models are all supported.
• Embedding model flexibility: choose from OpenAI, Cohere, Hugging Face, Google, Jina, or Ollama embeddings.
• Built on a familiar stack: React and Ant Design on the frontend, Node.js and Fastify on the backend, with PostgreSQL handling vector search and storage alongside Redis.
• Multiple interaction channels: a stable web embed script, plus beta Telegram and Discord bots and an experimental WhatsApp integration.
• Ships with a one-click Railway template and a standard Docker path for self-hosting.

Caveats

• The README carries an explicit disclaimer: this is a side project, not production-ready, and may contain bugs, security issues, or sudden breaking changes.
• Notion support and a Slack integration are on the roadmap but not yet implemented; several other integrations remain in beta or experimental stages.

Verdict Try it if you want a self-hosted knowledge-base chatbot that ingests everything from PDFs to YouTube links and swaps between OpenAI, Ollama, or Fireworks models on a whim. Skip it if you need a vendor-backed, production-hardened platform with a stability guarantee.