LM Studio Acquires Locally AI: Local LLMs Coming to iPhone and iPad

Local AI Goes Mobile: LM Studio Acquires Locally AI

On April 8, 2026, LM Studio — the leading desktop application for running large language models locally — announced the acquisition of Locally AI, a popular mobile app that brings on-device AI to iPhone, iPad, and Mac. The move marks a decisive strategic shift for LM Studio: from a desktop-first product to a true cross-device local AI platform. For millions of privacy-conscious users and developers who refuse to send their data to cloud services, this is one of the most significant expansions in the local AI ecosystem in recent memory.

What LM Studio and Locally AI Each Bring to the Table

LM Studio: The Desktop Standard

LM Studio launched in 2023 and quickly became the go-to graphical interface for downloading and running open-source LLMs on personal computers. Available on macOS, Windows, and Linux, it supports a broad catalog of models — including Qwen3, Gemma 3, DeepSeek-R1, Llama 4, and many others — sourced directly from the Hugging Face hub. Key capabilities include:

• OpenAI-compatible local API: Developers can point any OpenAI SDK client at localhost and use their chosen model without a single API key or cloud dependency.
• LM Link: A secure, end-to-end encrypted mesh VPN (built on Tailscale) that connects multiple LM Studio instances across devices, allowing a model running on a desktop to be accessed from another machine on the same network or remotely.
• llmster: A recently introduced headless mode that strips out the GUI, enabling LM Studio's inference core to run in server environments, CI/CD pipelines, and cloud VMs.
• JavaScript and Python SDKs: Programmatic access for building AI-powered applications that run entirely on the user's own hardware.

The company generated $1.8 million in revenue as of June 2025 with a 16-person team — lean and focused.

Locally AI: Mobile-Native On-Device Models

Built by developer Adrien Grondin, Locally AI earned a reputation as the most polished way to run personal AI models on Apple hardware. Unlike cloud-dependent iOS AI apps, Locally AI ran models entirely on-device — no server, no subscription, no data leaving the phone. For iPhone users who wanted the privacy guarantees of local AI without the friction of setting up a desktop server, Locally AI was the answer.

Grondin and the Locally AI product join LM Studio as the company accelerates its push into native mobile experiences.

Why This Acquisition Matters

The Gap in Local AI Has Always Been Mobile

Local AI on the desktop has been a solved problem for two-plus years. Tools like LM Studio, Ollama, and Jan.ai have made running a 7B or 13B parameter model on a laptop straightforward. But the mobile piece remained fragmented. Most approaches required the phone to connect to a desktop running a local server — workable, but clunky. Locally AI demonstrated that models could run natively on Apple Silicon iPhones and iPads, even if at reduced context and speed relative to desktop hardware.

With Apple Silicon M4 chips appearing in iPads and iPhones delivering dramatically improved neural engine throughput, the timing is right. Lightweight models in the 1B–3B range (like SmolLM2, Qwen2.5-0.5B, and Apple's own on-device models) already run comfortably on modern iPhones.

A Unified Local AI Ecosystem

LM Studio's stated goal after the acquisition is to build "new ways to use your models and agents seamlessly across your own devices." Combined with LM Link's existing cross-device mesh capabilities, the vision becomes clear: a user's models, conversations, and agent workflows could roam between desktop, laptop, iPhone, and iPad — all encrypted, all on-premises, never touching a third-party cloud.

This is an ambition that no major cloud AI provider can match by design. It is structurally differentiated.

Feature Roadmap Implications

While LM Studio has not published a detailed roadmap post-acquisition, several directions are strongly implied:

1. Native iOS and iPadOS app: Locally AI's existing iOS codebase and App Store presence provide the foundation. Expect a rebranded or integrated LM Studio mobile app.
2. Model sync across devices: Download a model once on desktop; access it on mobile via LM Link or direct transfer.
3. Cross-device agent workflows: An agent running a multi-step task could hand off between devices based on availability and compute capacity.
4. Deeper Apple Silicon optimization: Grondin's experience with Apple MLX framework and Core ML could accelerate LM Studio's existing MLX model support.

Usability Analysis

For current LM Studio desktop users, the acquisition does not change anything immediately. LM Studio 0.4.10 is the current version, and the mobile integration is still in development. However, users already benefit from LM Link for remote access to their desktop models from phones via browser — this acquisition promises to replace that workaround with a proper native experience.

For developers building privacy-first AI applications, this is a compelling platform bet. The free pricing model (no subscription for any LM Studio features), the OpenAI-compatible API, and the cross-device ambition make LM Studio an increasingly serious alternative to cloud inference for use cases where data locality is non-negotiable: legal, healthcare, personal journaling, and enterprise internal tools.

Pros and Cons

Pros

• First credible path to seamless local AI across desktop and Apple mobile devices
• Builds on an already best-in-class desktop experience
• Entirely free — no subscription fees, no usage metering
• Strong privacy guarantees: models and data never leave user hardware
• Adrien Grondin's deep Apple platform expertise strengthens the team

Cons

• Mobile features are not yet shipped; timeline unconfirmed
• Apple Silicon iPhones handle only small models (1B–3B parameters) at practical speeds
• Android support not mentioned — potentially Apple-only mobile strategy for now
• Small team (16 people) taking on significant cross-platform engineering scope

Outlook

The acquisition of Locally AI is the most concrete signal yet that the local AI movement is growing beyond its desktop roots. As on-device hardware from Apple, Qualcomm, and MediaTek continues to improve, the ceiling for mobile local AI rises every year. LM Studio is positioning itself to own this category before the cloud providers can credibly respond.

For enterprise IT teams increasingly concerned about data residency and AI governance, a mature cross-device local AI platform would address a real gap. LM Studio's free pricing removes budget friction, and open-source model compatibility means no vendor lock-in on the model side either.

Watch for a native iOS/iPadOS app announcement later in 2026 as the first concrete deliverable from this acquisition.

Conclusion

LM Studio's acquisition of Locally AI is a focused, strategic move that addresses the most obvious limitation of the local AI desktop ecosystem: the lack of a polished mobile experience. By bringing Adrien Grondin and the Locally AI codebase in-house, LM Studio signals a serious commitment to cross-device, privacy-first AI that runs entirely on user hardware. Developers and privacy-conscious users who have been waiting for local AI to reach their phones should watch this space closely. Rating: 4/5.