aihappey-chat

Open-source AI chat client with streaming, tools, MCP, and rich message rendering.

Open the live app · Watch Streaming UI demo · Storybook Chat · Storybook Themes

What this app is

aihappey-chat is an end-user AI chat client.
It helps you chat with models, switch providers, use tools/MCP, and work with rich outputs in one interface.

What you can do in this chat client

• Run normal chat conversations and see streamed replies in real time.
• Switch models and providers without leaving the same chat experience.
• Explore available models and compare options in the model explorer.
• Use the AI mesh view to understand provider/model relationships.
• Connect MCP servers and use MCP tools/resources during conversations.
• Use tool calling flows and inspect tool output directly in chat.
• Use Streaming UI apps/cards that appear during generation for dashboards, forms, and other structured outputs.
• Attach files and view rich message content (markdown, code, charts, PDFs, images, 3D).
• Use speech and transcription-related capabilities when enabled by your backend setup.
• Use reranking flows when you need document relevance scoring.
• Work with user settings and theme options (Fluent / Bootstrap).

Screenshots

1. AI model explorer: browse/filter models by modality and capabilities.
2. AI mesh: inspect how providers and models are connected.
3. Streaming UI: see response parts and updates as they arrive.

Supported Providers

The provider list is sourced from PROVIDERS and catalog definitions in packages/aihappey-core/src/runtime/providers/catalog.

Alphabetical ordered with links

302AI · Abliteration · AI21 · AIML · Alibaba · AmazonBedrock · Anthropic · APIpie · ARKLabs · ASIOne
AssemblyAI · AsyncAI · AtlasCloud · Audixa · Azure · Baseten · BergetAI · Bineric · Bria · BytePlus
Bytez · CanopyWave · Cerebras · Cirrascale · CloudRift · Cohere · CometAPI · ContextualAI · Cortecs · DeAPI
Decart · Deepbricks · Deepgram · DeepInfra · DeepL · DeepSeek · DigitalOcean · Echo · ElevenLabs · Euqai
EUrouter · Fireworks · Freepik · Friendli · Gladia · GMICloud · Google · GoogleTranslate · GooseAI · GreenPT
Groq · GTranslate · Helicone · HorayAI · Hyperbolic · Hyperstack · Inferencenet · Infomaniak · Inworld · IONOS
Jina · JSON2Video · KernelMemory · Kilo · KlingAI · LectoAI · Lingvanex · LLMGateway · MancerAI · MatterAI ·
MegaNova · MiniMax · Mistral · ModernMT · Moonshot · MurfAI · Nebius · Nextbit · NLPCloud · NousResearch
Novita · Nscale · NVIDIA · OpenAI · OpenRouter · OpperAI · OVHcloud · ParalonCloud · Parasail · Perplexity
Pollinations · Portkey · PrimeIntellect · PublicAI · Recraft · RegoloAI · RekaAI · RelaxAI · Replicate · Requesty
ResembleAI · Reve · Runpod · Runware · Runway · SambaNova · Sarvam · Scaleway · SEALION · Segmind
SiliconFlow · Speechify · Speechmatics · StabilityAI · StepFun · Sudo · SunoAPI · SUPA · Synexa · Telnyx
TencentHunyuan · Tinfoil · Together · TTSReader · Upstage · Verda · Vidu · VoyageAI · xAI · Zai

Getting started (optional, for local/self-hosted use)

1. Clone and install:

   git clone https://github.com/achappey/aihappey-chat.git
   cd aihappey-chat
   npm install
   

2. Configure environment:

   cd samples/chathappey
   copy .env.example .env
   

3. Run the sample app:

   npm run dev
   

4. In a second terminal:

   cd samples/chathappey
   npm run serve
   

See samples/chathappey/.env.example for all available variables.

Backend compatibility

This repository is client-side only and expects compatible backends:

• Vercel AI SDK compatible streaming chat backend (POST /api/chat)
• Vercel AI SDK compatible streaming agent backend (optional)
• Backend for MCP sampling calls
• MCP server for chat functionality

Optional backends:

• Remote conversation storage backend
• Transcription backend
• Entra ID authentication backend

Configuration

Key .env variables:

• CHAT_API_URL: chat backend endpoint (Vercel AI SDK compatible)
• AGENT_ENDPOINT: agent backend base URL (UI calls ${AGENT_ENDPOINT}/api/chat in agent mode)
• CHAT_APP_MCP: MCP server URL
• MODELS_API_URL: model catalog endpoint
• SAMPLING_API_URL: backend for MCP sampling calls (optional)
• CONVERSATIONS_API_URL: remote conversation storage backend (optional)
• TRANSCRIPTION_API: transcription backend (optional)

See samples/chathappey/.env.example for the complete list.

Developer notes

• Core chat integration is built around Vercel AI SDK primitives.
• useChat is re-exported via packages/aihappey-ai/src/index.ts.
• Main wiring happens in packages/aihappey-core/src/features/chat/engine/VercelChatInner.tsx.

Monorepo highlights:

• packages/aihappey-core: runtime logic and rich content
• packages/aihappey-components: reusable UI components
• packages/aihappey-mcp: MCP client
• samples/chathappey: reference app
• samples/bootstrap-sample: Bootstrap-flavored sample