Expanding Jibun Corp's AI Hub to 25 Providers

What Changed

The ai-hub Supabase Edge Function now supports 25 AI providers, up from 20. Five new additions:

The OpenAI-Compatibility Pattern Keeps Paying Off

24 of 25 providers use the same OPENAI_COMPAT_BODY function. Adding a new compatible provider is ~7 lines:

huggingface: {
  displayName: "Hugging Face",
  envKey: "HUGGINGFACE_TOKEN",
  chatUrl: "https://api-inference.huggingface.co/v1/chat/completions",
  defaultModel: "meta-llama/Llama-3.3-70B-Instruct",
  buildBody: OPENAI_COMPAT_BODY,
  parseResponse: OPENAI_COMPAT_PARSE,
},

The pattern reduces provider addition to a config entry — no custom parsing code needed.

Why Hugging Face Is Special

HuggingFace's Inference API gives access to 500,000+ models with a single API key. By pointing defaultModel at any model ID, users can effectively swap models without code changes. It's a shortcut to every open model in existence.

Why MiniMax Stands Out

MiniMax (香港上場・HKEX listed) goes beyond text: their platform integrates speech synthesis, video generation, and music generation. Text-01 supports up to 1 million tokens of context. The OpenAI-compatible endpoint makes it a drop-in replacement for any workflow.

Provider Status in Flutter UI

Each provider maps to a status in AiProviderStatus:

• implemented — API key set, ready to use
• apiKeyRequired — Code ready, Supabase Secret pending
• paidPlanRequired — Needs paid plan

All 5 new providers are apiKeyRequired — implementation complete, keys pending.

What's Next

The roadmap targets 80 providers for Phase 4. Next candidates include Reka AI, Replicate, IBM WatsonX, and Adept — each requiring API format verification before adding.

Building in public: https://my-web-app-b67f4.web.app/

FlutterWeb #Supabase #buildinpublic #AI