🧠Local LLM Runtimes arenaLocal LLM Runtimes

Agent access

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🤖Agenticness — how well agents can access and operate the productPoint an agent at llms.txt or agent-oriented docs	ai-native	✓fullT 8/10	—none 0/10	—none 0/10	✓fullT 8/10	—none 0/10	—none 0/10	~partialT 4/10
🤖Agenticness — how well agents can access and operate the productRun the product headlessly / in CI for automation	ai-native	~partialX 6/10	~partialC 6/10	~partialC 6/10	✓fullX 8/10	—none 0/10	~partialC 7/10	~partialX 6/10
🤖Agenticness — how well agents can access and operate the productPlug MCP servers into this product so it can use their tools	ai-native	—none 0/10	—none 0/10	n/a	~partialX 6/10	~partialC 6/10	✓fullC 8/10	n/a
🤖Agenticness — how well agents can access and operate the productConnect an agent via an official MCP server	ai-native	—none 0/10	—none 0/10	—none 0/10	n/a	n/a	✓fullC 7/10	n/a
🤖Agenticness — how well agents can access and operate the productUse an official CLI	ai-native	✓fullX 8/10	✓fullX 8/10	—none 0/10	✓fullX 8/10	—none 0/10	✓fullC 8/10	✓fullT 7/10
🤖Agenticness — how well agents can access and operate the productDrive the product through a documented public API	ai-native	✓fullT 8/10	~partialC 4/10	✓fullX 8/10	✓fullX 8/10	~partialT 5/10	✓fullT 8/10	~partialT 5/10
🤖Agenticness — how well agents can access and operate the productIssue scoped/least-privilege API credentials for an agent	ai-native	—none 0/10	n/a	—none 0/10	n/a	—none 0/10	~partialC 3/10	n/a
🤖Agenticness — how well agents can access and operate the productBuild against official SDKs	ai-native	✓fullC 7/10	—none 0/10	~partialX 5/10	—none 0/10	—none 0/10	~partialT 6/10	—none 0/10
🤖Agenticness — how well agents can access and operate the productSubscribe to events via webhooks	ai-native	—none 0/10	—none 0/10	n/a	n/a	—none 0/10	—none 0/10	n/a
🤖Agenticness — how well agents can access and operate the productConnect a coding agent to this product as a working backend	ai-native	✓fullC 8/10	~partialC 4/10	~partialX 6/10	~partialX 7/10	~partialT 6/10	✓fullC 8/10	~partialT 4/10

Agentic features

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🤖Agenticness — how well agents can access and operate the productGet AI-generated insights and suggestions from my data inside the product	ai-native	n/a	—none 0/10	n/a	~partialX 6/10	—none 0/10	~partialC 5/10	—none 0/10
🤖Agenticness — how well agents can access and operate the productSet up automations that run autonomously in the background	ai-native	n/a	—none 0/10	n/a	—none 0/10	—none 0/10	~partialC 4/10	n/a
🤖Agenticness — how well agents can access and operate the productDelegate tasks to a built-in AI assistant inside the product	ai-native	—none 0/10	—none 0/10	n/a	~partialX 6/10	~partialC 6/10	✓fullC 8/10	—none 0/10
🤖Agenticness — how well agents can access and operate the productOperate the product with natural-language commands	ai-native	~partialC 5/10	—none 0/10	n/a	~partialX 6/10	~partialC 5/10	~partialC 6/10	~partialX 6/10

Api quality

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🤖Agenticness — how well agents can access and operate the productExplore an interactive API reference with runnable examples	ai-native	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
🤖Agenticness — how well agents can access and operate the productDownload a machine-readable API spec (OpenAPI or equivalent)	ai-native	—none 0/10	—none 0/10	—none 0/10	—none 0/10	~partialT 4/10	—none 0/10	—none 0/10
🤖Agenticness — how well agents can access and operate the productTest against a sandbox environment without touching production data	ai-native	n/a	n/a	n/a	—none 0/10	n/a	—none 0/10	n/a
🤖Agenticness — how well agents can access and operate the productRely on versioned APIs with a documented deprecation policy	ai-native	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Build and install

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesBuild the runtime from source with minimal external dependencies	developer	—none 0/10	✓fullX 8/10	~partialC 3/10	—none 0/10	~partialC 4/10	—none 0/10	—none 0/10
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesRun the runtime inside a container for reproducible deployment	developer	~partialX 3/10	✓fullC 8/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesInstall the runtime quickly using a standard package manager	developer	—none 0/10	~partialX 4/10	✓fullC 8/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesInstall using prebuilt binaries or packages instead of compiling from source	developer	~partialT 5/10	~partialX 6/10	✓fullC 7/10	✓fullX 7/10	—none 0/10	—none 0/10	✓fullX 8/10

Community contribution

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesContribute code and become a recognized collaborator through the project's open-source process	developer	—none 0/10	~partialC 5/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Language bindings

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesCall the runtime from official client libraries in languages like Python or JavaScript	developer	✓fullC 8/10	—none 0/10	~partialX 5/10	—none 0/10	~partialT 4/10	~partialT 6/10	—none 0/10

Licensing and cost

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesWhether commercial or enterprise use requires a paid license or subscription beyond the free community edition	power-user	—none 0/10	—none 0/10	n/a	—none 0/10	—none 0/10	n/a	n/a

Maintenance health

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesHow quickly the project ships patches for critical bugs and security vulnerabilities based on its public release history	developer	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Model portability

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesWhether downloaded model files and caches can be reused by other runtimes without re-downloading or re-converting them	developer	—none 0/10	—none 0/10	~partialC 3/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Privacy control

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Ecosystem — integrations, plugins, and third-party ecosystem storiesRun inference entirely on my own machine so my data and prompts never leave my device	power-user	✓fullT 8/10	✓fullX 9/10	✓fullX 8/10	✓fullX 9/10	✓fullX 8/10	✓fullC 9/10	✓fullX 9/10

Architecture coverage

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Model support — which models run and how well — coverage, formats, update cadenceRun hundreds of different model architectures including LLMs, MoE, multi-modal, and embedding models	developer	~partialT 5/10	~partialX 6/10	✓fullX 9/10	~partialX 5/10	~partialC 5/10	~partialC 6/10	~partialX 6/10
✨Model support — which models run and how well — coverage, formats, update cadenceServe embedding models for retrieval and search applications	developer	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Custom assistants

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Model support — which models run and how well — coverage, formats, update cadenceCreate specialized custom assistants configured for specific tasks	power-user	—none 0/10	~partialC 3/10	~partialC 5/10	~partialC 4/10	~partialC 5/10	✓fullC 8/10	~partialX 5/10

Hybrid cloud local

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Model support — which models run and how well — coverage, formats, update cadenceConnect to cloud AI providers alongside local models within the same interface	power-user	~partialT 6/10	n/a	n/a	—none 0/10	✓fullC 8/10	—none 0/10	—none 0/10
✨Model support — which models run and how well — coverage, formats, update cadenceOffload very large models to a hosted cloud tier without downloading them when my local hardware is insufficient	power-user	✓fullC 7/10	—none 0/10	n/a	—none 0/10	—none 0/10	—none 0/10	—none 0/10
✨Model support — which models run and how well — coverage, formats, update cadenceThe pricing tiers, rate limits, and usage caps that apply when offloading inference to the vendor's hosted cloud tier	power-user	—none 0/10	n/a	n/a	n/a	—none 0/10	n/a	n/a

Model hub download

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Model support — which models run and how well — coverage, formats, update cadenceDownload and run open models directly from Hugging Face	power-user	~partialX 5/10	✓fullX 8/10	✓fullX 8/10	✓fullX 8/10	✓fullC 8/10	✓fullC 9/10	—none 0/10

Multi modal support

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Model support — which models run and how well — coverage, formats, update cadenceRun vision-language models that understand images alongside text	power-user	~partialX 5/10	✓fullX 8/10	—none 0/10	—none 0/10	—none 0/10	~partialC 3/10	✓fullC 8/10

Distributed serving

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useDisaggregate prefill and decode phases for optimized large-scale serving	developer	n/a	—none 0/10	~partialC 6/10	n/a	n/a	—none 0/10	—none 0/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useDistribute inference across multiple GPUs using tensor, pipeline, or data parallelism	developer	—none 0/10	—none 0/10	✓fullC 8/10	—none 0/10	—none 0/10	~partialC 3/10	—none 0/10

Gpu acceleration

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useRun inference on specialized accelerators like TPUs or Gaudi through plugin support	developer	—none 0/10	—none 0/10	✓fullC 8/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useRun models larger than my available VRAM using combined CPU+GPU offload	power-user	—none 0/10	✓fullX 8/10	—none 0/10	~partialC 3/10	—none 0/10	—none 0/10	~partialX 5/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useWhy GPU acceleration failed and silently fell back to CPU through clear diagnostic output	power-user	~partialX 4/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	!disputedD 3/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useRun models on NVIDIA, AMD, or other GPU vendors using vendor-specific acceleration kernels	power-user	~partialX 5/10	✓fullX 8/10	✓fullC 8/10	~partialX 4/10	—none 0/10	~partialC 6/10	!disputedD 5/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useAccelerate inference on AMD GPUs via a Vulkan backend without needing a full ROCm install	power-user	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	~partialC 4/10

Memory management

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useControl how context memory is allocated when running multiple model instances concurrently	power-user	—none 0/10	—none 0/10	~partialX 6/10	~partialC 6/10	—none 0/10	—none 0/10	~partialX 2/10

Platform acceleration

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useGet accelerated inference on Apple Silicon via native ARM and Metal optimizations	power-user	~partialX 4/10	✓fullX 9/10	~partialC 3/10	~partialX 5/10	—none 0/10	—none 0/10	~partialX 5/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useRun inference on diverse CPU architectures beyond x86 and ARM, such as PowerPC	developer	—none 0/10	—none 0/10	✓fullC 8/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useLeverage advanced x86 CPU instruction sets like AVX, AVX2, AVX512, and AMX for faster inference	power-user	—none 0/10	✓fullX 8/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Startup footprint

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useGet a fast cold start from a lightweight runtime binary instead of waiting seconds before inference begins	power-user	—none 0/10	✓fullX 7/10	—none 0/10	!disputedD 4/10	—none 0/10	—none 0/10	~partialX 5/10

Throughput optimization

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useAchieve high serving throughput via continuous batching and chunked prefill	power-user	—none 0/10	—none 0/10	✓fullX 9/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useRely on paged memory management for attention key/value cache to maximize concurrent request capacity without memory fragmentation	developer	—none 0/10	—none 0/10	✓fullX 9/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useThe runtime reserves dedicated capacity so throughput holds steady when multiple agents or sessions issue requests concurrently	power-user	~partialC 4/10	—none 0/10	~partialX 5/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useSpeed up repeated-prompt workloads using prefix caching	power-user	—none 0/10	—none 0/10	~partialX 6/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
⚡Performance hardware — raw speed and hardware efficiency — throughput, latency, resource useAccelerate generation speed using speculative decoding techniques	power-user	—none 0/10	—none 0/10	~partialC 6/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Adapters

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Quantization formats — stories about quantization formats in this arenaEfficiently serve multiple LoRA adapters on top of a base model	developer	—none 0/10	—none 0/10	✓fullX 8/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

File formats

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Quantization formats — stories about quantization formats in this arenaWhether upgrading the runtime can break compatibility with previously downloaded quantized model files	developer	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
✨Quantization formats — stories about quantization formats in this arenaLoad and run models packaged in the GGUF format	power-user	—none 0/10	✓fullX 8/10	✓fullX 8/10	~partialX 5/10	~partialC 6/10	✓fullC 9/10	✓fullX 7/10

Quantization levels

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
✨Quantization formats — stories about quantization formats in this arenaReduce memory footprint using integer quantization ranging from very low-bit to 8-bit precision	power-user	~partialX 4/10	✓fullX 9/10	✓fullX 7/10	—none 0/10	—none 0/10	~partialC 3/10	—none 0/10
✨Quantization formats — stories about quantization formats in this arenaLoad models quantized in formats like FP8, INT4, GPTQ, or AWQ	developer	—none 0/10	—none 0/10	✓fullX 8/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Api compatibility

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityCall the server through an Anthropic-compatible messages endpoint	developer	—none 0/10	—none 0/10	~partialC 6/10	—none 0/10	—none 0/10	✓fullC 8/10	—none 0/10
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityLaunch a local OpenAI-compatible API server for any loaded model	developer	~partialT 6/10	~partialC 6/10	✓fullX 8/10	✓fullX 9/10	✓fullC 7/10	✓fullC 9/10	~partialC 5/10

Deployment modes

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityRun the runtime headlessly with no GUI for use in servers or CI pipelines	developer	~partialX 6/10	✓fullC 8/10	✓fullX 7/10	✓fullX 8/10	—none 0/10	✓fullC 7/10	✓fullX 8/10

Generation controls

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityStream generated tokens back to my application as they are produced	developer	—none 0/10	~partialX 4/10	✓fullC 7/10	~partialX 4/10	~partialT 5/10	~partialC 4/10	~partialC 4/10
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityConstrain model output to structured formats like JSON using grammars	developer	—none 0/10	✓fullC 7/10	~partialC 6/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityUse native tool-calling and reasoning-parser support in my requests	developer	—none 0/10	—none 0/10	~partialC 6/10	—none 0/10	—none 0/10	~partialC 6/10	—none 0/10

Model lifecycle

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityAssign a custom identifier to a loaded model for consistent reference in API calls	developer	—none 0/10	—none 0/10	—none 0/10	✓fullC 8/10	—none 0/10	✓fullC 7/10	—none 0/10
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityLoad and switch between multiple models without restarting the server	power-user	✓fullX 8/10	—none 0/10	~partialC 4/10	~partialC 6/10	~partialC 4/10	~partialC 5/10	—none 0/10

Remote serving

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityServe models over my local network for access from other devices	power-user	~partialX 4/10	~partialC 6/10	~partialX 5/10	~partialX 6/10	~partialC 5/10	~partialC 6/10	~partialC 5/10

Scale limits

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityThe documented maximum concurrent requests or connections the local server can handle before throughput degrades	developer	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Server configuration

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🔌Serving api — serving models over an API — endpoints, compatibility, reliabilityOverride low-level engine settings like memory locking or mmap behavior instead of being limited to opinionated defaults	power-user	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Ai assisted setup

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingRely on an AI assistant to recommend which local model best fits my hardware and task before I download it	ai-native	—none 0/10	—none 0/10	n/a	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Chat interface

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingChat with local models using a built-in graphical chat interface	power-user	~partialX 6/10	✓fullC 7/10	n/a	✓fullX 9/10	✓fullX 7/10	✓fullC 8/10	✓fullX 7/10

Cli tooling

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingStart an interactive chat session with a model directly from the terminal	developer	✓fullX 8/10	✓fullX 8/10	—none 0/10	✓fullC 8/10	—none 0/10	✓fullC 7/10	✓fullX 8/10
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingSearch, download, and manage models from a command-line interface	developer	✓fullX 8/10	~partialC 6/10	—none 0/10	✓fullC 8/10	—none 0/10	✓fullC 7/10	—none 0/10
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingLoad a model with custom GPU offload and context length settings from the command line	developer	—none 0/10	~partialX 6/10	—none 0/10	✓fullX 8/10	—none 0/10	—none 0/10	~partialT 6/10
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingStart and stop the local model server from the command line	developer	~partialX 3/10	~partialC 6/10	—none 0/10	✓fullC 9/10	—none 0/10	~partialC 5/10	~partialT 5/10
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingLaunch popular third-party coding agent CLIs pre-configured to use my local models with a single command	developer	✓fullC 7/10	—none 0/10	n/a	—none 0/10	—none 0/10	—none 0/10	—none 0/10

Document intelligence

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingChat with my own documents entirely offline using automatic retrieval-augmented generation	ai-native	—none 0/10	—none 0/10	n/a	✓fullX 7/10	—none 0/10	—none 0/10	—none 0/10
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingHave an AI agent draft and edit documents in an integrated workspace with changes saved automatically	ai-native	n/a	n/a	n/a	✓fullX 7/10	—none 0/10	—none 0/10	n/a
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingDictate speech that gets transcribed in real time by an on-device model	ai-native	n/a	n/a	n/a	✓fullC 5/10	—none 0/10	~partialC 5/10	~partialC 4/10

Local model management

Story	Persona	Ollama	llama.cpp	vLLM	LM Studio	Jan	LocalAI	llamafile
🎨Ux tooling — the working surface itself — layout, ergonomics, quality-of-life toolingManage my downloaded models, saved prompts, and per-model configurations in one place	power-user	~partialX 3/10	—none 0/10	n/a	~partialX 7/10	~partialC 4/10	~partialC 5/10	—none 0/10