Batuk — Sovereign AI for Teams

Run every AI model securely in one enterprise workspace

Developer Tools

Artificial Intelligence

GitHub

Visit WebsiteSee on Product HuntGithub ⧉

👋 Hi Product Hunt! I’m Suhas, the creator of Batuk. Over the past year, I kept seeing the same problem across AI projects: companies wanted to use different AI models, but they did not want their teams switching between disconnected chat tools, sharing provider credentials, uploading documents to unknown systems, or losing control over where their data was stored. That is why I built Batuk. Batuk is an open-source, sovereign AI workspace that gives teams one interface for the models they are permitted to use. It supports: 🤖 Ollama for local and offline models 🧠 OpenAI, Claude, Grok, OpenRouter, and Sarvam AI 🔌 Private OpenAI-compatible endpoints such as vLLM, LM Studio, llama.cpp, LiteLLM, and internal AI gateways 📚 Document chat and RAG across PDFs, Word files, spreadsheets, Markdown, CSV, JSON, logs, and text files 🔐 Email authentication, OAuth, OIDC, SAML SSO, SCIM provisioning, organizations, teams, roles, and invitations 🛡️ Admin controls for user management, password resets, access restrictions, and organization-level configuration 📋 Audit trails, hash-chain integrity evidence, GDPR request workflows, data inventories, retention notes, and ISO 27001/SOC 2 control registers 🎨 Organization-level whitelabeling with custom names, logos, taglines, initials, and accent colors 🧩 Reusable AI skills, agent workflows, document-backed agents, web search, voice sessions, guardrails, and temporary chats 🐳 Docker-based deployment for private-cloud, on-premises, and air-gapped environments Batuk is local-first by default. A developer can run it on a laptop using Ollama, SQLite, local files, and JSON-based vector storage. A company can deploy the same platform using PostgreSQL or MySQL, ChromaDB or Pinecone, private document storage, enterprise identity providers, and organization-scoped data. The goal is not to create another wrapper around a single model. The goal is to provide a practical foundation that companies can own, inspect, customize, deploy, and operate within their own security boundaries. Batuk is particularly useful for: • Companies evaluating sovereign or private AI • Internal AI platform teams • Regulated organizations • Consulting firms deploying AI environments for clients • Teams using both cloud and local models • Organizations that need SSO, auditability, data isolation, and deployment control before adopting AI internally The entire project is open source under the MIT License. This is an early launch, and I would genuinely value feedback from CTOs, security leaders, AI engineers, enterprise architects, and teams already running private LLM infrastructure. I would especially love to hear: 1. Which model providers or private gateways should be supported next? 2. Which enterprise controls would you require before deploying Batuk internally? 3. Would you use Batuk as an internal AI workspace, a client deployment foundation, or both? 4. Which integrations would make it immediately useful for your team? Thank you for checking out Batuk. Your feedback will help shape what I build next. 🙏