Eron v1.4: $2.99 native iOS client for Ollama with zero-buffer streaming and HomeKit tools
RA2B_DIN · reddit · 2026-10-09
Developer RA2BDIN released Eron v1.4, a native iOS client for Ollama, vLLM, and LM Studio users (or BYOK), priced as a flat $2.99 one-time purchase — a rebuttal to App Store clients that charge $15/month and route prompts through cloud proxies.
Key architecture points:
- Zero proxy: direct HTTP/WebSocket connection to local IP or Tailscale/WireGuard nodes; no telemetry, no account
- Zero-buffer streaming: rewritten pipeline renders raw token chunks as fast as the GPU produces them
- Reasoning stream: native collapsible rendering of <think> blocks (DeepSeek R1, Qwen reasoning models)
- Local tool calling: bridges to Apple Reminders, Calendar, and HomeKit smart home when the model supports function calling
- Workspaces: isolated projects with persistent custom system prompts
- v1.4.1 adds a native dual-screen layout for the rumored iPhone Duo form factor
The dev is giving away 20 promo codes for local-setup users.
More from Infra
- Hadfield-Menell: We should track capability cost and access, not just open weights — dhadfieldmenell · 2026-10-09
- $2,800 rig of 8x Radeon Pro V620 (256GB VRAM) hits 3000+ t/s prefill via custom vLLM fork — _TheWolfOfWalmart_ · 2026-10-09
- Coatue lays out new forms of financing for AI compute — _AustinCalvert_ · 2026-10-09
- Benchmarking 4 open decision models on one RTX 4090: Laya fastest, Lev most accurate at 13x the latency — Fun-Meaning-6474 · 2026-10-09
- fal Engineer on Video Speed: H3 Max Renders 15 Seconds of Video in 5 — OdinLovis · 2026-10-09
- Microsoft and NVIDIA Going Hard on Local AI as a Defense Against Frontier Labs — MatthewBerman · 2026-10-09