--- title: Testing Local LLMs Used to Be a Chore. Crustly’s New... description: Crustly v0.5.2 is a Rust-based terminal AI assistant that now integrates natively with Ollama, enabling developers to download, switch, and tune local LLMs... canonical: https://daily.dev/posts/testing-local-llms-used-to-be-a-chore-crustly-s-new-ollama-integration-changes-everything--n93qkerhw twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Testing Local LLMs Used to Be a Chore. Crustly’s New Ollama Integration Changes Everything. | daily.dev og:description: Crustly v0.5.2 is a Rust-based terminal AI assistant that now integrates natively with Ollama, enabling developers to download, switch, and tune local LLMs... og:url: https://daily.dev/posts/testing-local-llms-used-to-be-a-chore-crustly-s-new-ollama-integration-changes-everything--n93qkerhw og:image: https://api.daily.dev/og/posts/N93QKeRhw.png og:image:alt: Testing Local LLMs Used to Be a Chore. Crustly’s New Ollama Integration Changes Everything. og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Testing Local LLMs Used to Be a Chore. Crustly’s New Ollama Integration Changes Everything. **[Medium](https://daily.dev/sources/medium_js)** · 7 min read · 0 upvotes · 0 comments ## Summary Crustly v0.5.2 is a Rust-based terminal AI assistant that now integrates natively with Ollama, enabling developers to download, switch, and tune local LLMs entirely within a TUI. Key features include Ctrl+D to pull models with a live progress bar, Ctrl+W to switch models mid-session while preserving per-model configuration, real-time performance metrics (tokens/sec, generation time, cold/warm start), and a 22-tool agentic system for file operations and shell commands with approval gating. Per-model config overrides let users set context window size and temperature independently for each model. Version 0.5.2 also fixes several reliability bugs including a database race condition, a false-positive loop detector, and broken config persistence on model switch. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How do I switch between different Ollama models mid-conversation in Crustly without losing my configuration? Press Ctrl+W in Crustly, pick any installed model, and press Enter — the next message runs on that model. As of version 0.5.2, the switch carries the model's full configuration with it, including per-model overrides like context window and temperature, so a switched-to model behaves exactly as if you had started the session with it. _daily.dev surfaces practical workflow notes like this for developers comparing local LLM tooling._ ### Why does Crustly show tokens per second and cold start time under each response when using Ollama? Crustly's native Ollama provider speaks Ollama's own /api/chat protocol rather than the OpenAI compatibility shim, which lets it report tokens per second, generation time, and whether the reply was a warm or cold start under every response. This gives real performance data for comparing local models on your own hardware rather than relying on estimates. _developers benchmarking local models on their own hardware can find breakdowns like this on daily.dev._ ### What bugs did Crustly version 0.5.2 fix for its agentic terminal AI assistant? Version 0.5.2 fixed a database race that crashed every non-interactive crustly run invocation with a Message not found error, a false-positive loop detector that silently aborted legitimate consecutive file edits, Ctrl+W dropping configuration when rebuilding the provider, and tool calls printed as raw JSON text by some model templates not being recovered and executed. _track reliability fixes like these for agentic coding tools on daily.dev before upgrading._ ## Similar posts on daily.dev - [Ollama is the easiest way to start local LLMs, but these 6 alternatives are also worth trying](https://daily.dev/posts/ollama-is-the-easiest-way-to-start-local-llms-but-these-6-alternatives-are-also-worth-trying-qmo5um93q) · XDA Developers · 0 upvotes · 0 comments - [Don’t Use Ollama for Local LLMs](https://daily.dev/posts/don-t-use-ollama-for-local-llms-szl4snpls) · Medium · 0 upvotes · 0 comments - [Everything you need to know about running LLMs locally](https://daily.dev/posts/everything-you-need-to-know-about-running-llms-locally-7u3awxooq) · All Things Open · 0 upvotes · 0 comments - [Supercharging GitHub Copilot CLI with Ollama: Local Models, Full Control](https://daily.dev/posts/supercharging-github-copilot-cli-with-ollama-local-models-full-control-assgnjkiq) · The Art of Simplicity · 0 upvotes · 0 comments - [How to Run and Customize LLMs Locally with Ollama](https://daily.dev/posts/how-to-run-and-customize-llms-locally-with-ollama-cixkmhbqh) · freeCodeCamp · 2 upvotes · 0 comments --- Tags: [#rust](https://daily.dev/tags/rust), [#local-ai](https://daily.dev/tags/local-ai), [#ollama](https://daily.dev/tags/ollama) [View this post on daily.dev](https://daily.dev/posts/testing-local-llms-used-to-be-a-chore-crustly-s-new-ollama-integration-changes-everything--n93qkerhw) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Testing Local LLMs Used to Be a Chore. Crustly’s New Ollama Integration Changes Everything.","url":"https://daily.dev/posts/testing-local-llms-used-to-be-a-chore-crustly-s-new-ollama-integration-changes-everything--n93qkerhw","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/testing-local-llms-used-to-be-a-chore-crustly-s-new-ollama-integration-changes-everything--n93qkerhw"},"datePublished":"2026-07-23T07:30:35.616Z","dateModified":"2026-09-13T19:15:56.670Z","description":"Crustly v0.5.2 is a Rust-based terminal AI assistant that now integrates natively with Ollama, enabling developers to download, switch, and tune local LLMs...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/75fa9910d697f480b041815cb9582f9b?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/75fa9910d697f480b041815cb9582f9b?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Medium","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Medium","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/medium","url":"https://daily.dev/sources/medium_js"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/testing-local-llms-used-to-be-a-chore-crustly-s-new-ollama-integration-changes-everything--n93qkerhw","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"rust,local-ai,ollama","timeRequired":"PT7M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Medium","item":"https://daily.dev/sources/medium_js"},{"@type":"ListItem","position":3,"name":"Testing Local LLMs Used to Be a Chore. Crustly’s New Ollama Integration Changes Everything."}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/testing-local-llms-used-to-be-a-chore-crustly-s-new-ollama-integration-changes-everything--n93qkerhw#faq","mainEntity":[{"@type":"Question","name":"How do I switch between different Ollama models mid-conversation in Crustly without losing my configuration?","acceptedAnswer":{"@type":"Answer","text":"Press Ctrl+W in Crustly, pick any installed model, and press Enter — the next message runs on that model. As of version 0.5.2, the switch carries the model's full configuration with it, including per-model overrides like context window and temperature, so a switched-to model behaves exactly as if you had started the session with it. daily.dev surfaces practical workflow notes like this for developers comparing local LLM tooling."}},{"@type":"Question","name":"Why does Crustly show tokens per second and cold start time under each response when using Ollama?","acceptedAnswer":{"@type":"Answer","text":"Crustly's native Ollama provider speaks Ollama's own /api/chat protocol rather than the OpenAI compatibility shim, which lets it report tokens per second, generation time, and whether the reply was a warm or cold start under every response. This gives real performance data for comparing local models on your own hardware rather than relying on estimates. developers benchmarking local models on their own hardware can find breakdowns like this on daily.dev."}},{"@type":"Question","name":"What bugs did Crustly version 0.5.2 fix for its agentic terminal AI assistant?","acceptedAnswer":{"@type":"Answer","text":"Version 0.5.2 fixed a database race that crashed every non-interactive crustly run invocation with a Message not found error, a false-positive loop detector that silently aborted legitimate consecutive file edits, Ctrl+W dropping configuration when rebuilding the provider, and tool calls printed as raw JSON text by some model templates not being recovered and executed. track reliability fixes like these for agentic coding tools on daily.dev before upgrading."}}]} ```