AI tool comparison
LM Studio 0.4.0 vs Stash
Which one should you ship with? Here is the side-by-side panel verdict, pricing read, reviewer split, and community vote comparison.
Local AI Infrastructure
LM Studio 0.4.0
Local LLMs get a headless CLI — run models as a server daemon anywhere
100%
Panel ship
—
Community
Free
Entry
LM Studio 0.4.0 is the biggest update to the popular local LLM runner since its launch, introducing a proper headless CLI that separates the model inference engine from the GUI entirely. The new `lms` / `llmster` command starts LM Studio as a daemon — no display required — making local models viable in CI pipelines, remote servers, Docker containers, and scheduled tasks for the first time. The update ships three major features alongside the CLI: continuous batching for parallel requests (multiple simultaneous users against one running model), a stateful `/v1/chat` REST API that preserves conversation state across calls without the client managing message history, and an interactive terminal chat via `lms chat` with streaming and system prompt support. The headless mode pairs naturally with Claude Code via a `claude-lm` alias that routes Claude's tool calls to the local model. LM Studio 0.4.0 landed on Hacker News with 216 points, driven heavily by the "Running Gemma 4 locally" angle — Gemma 4's efficiency makes it one of the best models to run under 0.4.0's new architecture. The stateful API is particularly notable: it means the inference server maintains context between API calls, which dramatically simplifies agent loop implementations that don't want to re-send full conversation history on every turn.
Infrastructure
Stash
Open-source memory layer that teaches AI agents to remember and learn
75%
Panel ship
—
Community
Paid
Entry
Stash is an open-source persistent memory infrastructure for AI agents built on PostgreSQL and pgvector. Unlike retrieval-augmented generation, which searches static documents, Stash actively learns from agent experience — consolidating raw observations into facts, relationships, causal links, and higher-order patterns over time. The system exposes 28 MCP tools covering the full cognitive stack: episode storage, fact synthesis, entity graph management, goal tracking, failure pattern recognition, and self-correction when contradictions emerge. It deploys via Docker Compose in three steps and works with any OpenAI-compatible API — Claude, GPT, local models via Ollama. Hierarchical namespaces let agents keep user facts separate from project facts separate from self-knowledge. This fills a real gap in the agent ecosystem. Most agent frameworks treat each session as stateless, which means agents repeat the same mistakes and lose hard-won context. Stash gives agents a persistent cognitive layer that compounds. It surfaced on Hacker News this week to notable developer interest and is worth watching as MCP adoption accelerates.
Reviewer scorecard
“The headless CLI and stateful /v1/chat API are the two things keeping LM Studio off my production stack. With 0.4.0, I can finally run local models in CI and point agents at them without managing conversation state on the client. This is the version I've been waiting for.”
“The 28 MCP tools are the right abstraction level — my Claude Desktop agents can now actually remember what I've told them across sessions without me writing my own memory layer. The Docker Compose setup is clean and the pgvector backend is production-ready.”
“I'm skeptical of local LLM tooling that ships half-finished features, but the headless CLI is genuinely production-ready based on early reports. My only concern: continuous batching on consumer hardware degrades quality under load. Test your specific hardware before committing.”
“The consolidation pipeline sounds elegant in theory but in practice you're letting an LLM synthesize 'causal links' and 'higher-order patterns' from raw observations. That's a recipe for hallucinated beliefs that compound over time. I'd want rigorous testing before trusting this in any production agent.”
“LM Studio going headless is a pivotal moment for local AI infrastructure. When you can run a fully capable local model as a daemon with a stateful REST API, the cloud API becomes optional for the majority of use cases. The cost and privacy implications are enormous.”
“Persistent memory is the missing piece between 'AI assistant' and 'AI colleague.' Stash's self-correction and failure pattern recognition are early implementations of what agents will need to become genuinely reliable over long time horizons.”
“I'm not a developer but I run LM Studio for private writing and research. The new terminal chat is cleaner than the GUI for long sessions, and knowing it runs as a background daemon means I can finally build simple automations on top of my local models.”
“Finally an agent that remembers my brand guidelines, tone preferences, and past feedback without me repeating myself every session. The namespace hierarchy means I can have separate memories for different clients.”
Weekly AI Tool Verdicts
Get the next comparison in your inbox
New AI tools ship daily. We compare them before you waste an afternoon.