Comments
>log in to commentUnlike cobbling together separate AI services, hfy gives you a single backend and chat UI you can run next to your app. It handles the model, tool calls, history, and UI while keeping your users and data in your own database. I built it to stop teams from re‑implementing the same plumbing. Let me know what works and what could be better 😊
Self-hosted backend and embeddable chat UI that lets your app run AI assistants with your own data and tools.
- for
- Product teams that need an on-prem AI chat for web, mobile or desktop apps.
- pricing
- open source
- license
- MIT
Key features
- Self-hosted deployment — Runs on your own servers, keeping user data and API keys in-house.
- Model agnostic — Works with OpenAI, OpenRouter, local models or any service using the OpenAI API format.
- Tool integration via MCP — Calls your internal APIs through the open MCP standard, with per-request identity.
- Persistent conversations — All messages are stored in Postgres, surviving refreshes, crashes and device switches.
- Native chat UI — Embeddable web, iOS, Android, Flutter or React-Native chat that matches your brand.
- Memory & summarization — Long dialogs are summarized and searchable so the assistant can recall past context.
- Confirmation flows — Risky actions like refunds wait for explicit user approval before execution.
- Multi-assistant rooms — Separate assistants for customers and staff can share a conversation with role-based views.
Use cases
- Customer support chat that can look up orders, issue refunds and attach photos.
- E-commerce assistant that searches product catalogs, suggests items and handles size/budget questions.
- Team copilot that accesses internal tools, runs reports and performs admin tasks on behalf of staff.
hfy vs alternatives
 hfy |  n8n |  LangChain |  LlamaIndex |  CrewAI | |
|---|---|---|---|---|---|
| Best for | Self-hosted, native-look AI chat for apps | Workflow automation with AI | LLM app framework | Data indexing for LLMs | Multi-agent AI orchestration |
| Pricing | Open source | Subscription | Free | Free | Subscription |
| DevHunt upvotes | 1 | 0 | 0 | 0 | 0 |
| Launched | Oct 2026 | — | — | — | — |
- hfy vs n8n: n8n is a general workflow automation platform, not a dedicated embeddable chat UI.
- hfy vs LangChain: LangChain is a developer framework for building LLM apps, whereas hfy provides a ready-made backend and UI.
- hfy vs LlamaIndex: LlamaIndex focuses on indexing and querying data for LLMs, not on chat UI or self-hosted assistant services.
- hfy vs CrewAI: CrewAI orchestrates multi-agent systems, while hfy delivers a single-assistant chat experience embedded in your product.
hfy FAQ
Do I need to host hfy myself?+
Yes, hfy is not a hosted service; you run it on your own infrastructure and store conversations in your database.
Can I use any LLM with hfy?+
Any model that speaks the OpenAI API format can be used, including OpenAI, OpenRouter, or self-hosted models.
How does hfy handle risky actions like refunds?+
Actions marked as risky require explicit user confirmation before the assistant executes them.
Is there an SDK to embed the chat?+
No SDK is needed; you embed the provided chat UI component and configure assistants via a settings file.
What storage does hfy require?+
A PostgreSQL database is required for persisting conversations, files and metadata.
Is the code open source?+
Yes, hfy is MIT-licensed and the source is publicly available.
Summarized by DevHunt from alash3al.github.io · Oct 27, 2026. Details may change; check the official site.
-1200%D1%851200-px.png?width=64)
