--- title: Lumen 1.4.0 Release | daily.dev description: Lumen 1.4 makes the conversational data analysis framework easier to start and cheaper to run. You can now use it without an API key via Claude Code, Codex,... canonical: https://daily.dev/posts/lumen-1-4-0-release-oivtn6yi5 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Lumen 1.4.0 Release | daily.dev og:description: Lumen 1.4 makes the conversational data analysis framework easier to start and cheaper to run. You can now use it without an API key via Claude Code, Codex,... og:url: https://daily.dev/posts/lumen-1-4-0-release-oivtn6yi5 og:image: https://api.daily.dev/og/posts/oIVtn6YI5.png og:image:alt: Lumen 1.4.0 Release 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. # Lumen 1.4.0 Release **[Planet Python](https://daily.dev/sources/planetpython)** · 18 min read · 0 upvotes · 0 comments ## Summary Lumen 1.4 makes the conversational data analysis framework easier to start and cheaper to run. You can now use it without an API key via Claude Code, Codex, GitHub Copilot CLI, Kilo's free tier, or a local Ollama model. Tables with geometry columns render as choropleth maps, reports can get an AI-written narrative and export to Word, and every SQL result gets automatic data-quality checks. Table context for agents shrank from 22.4k to 16.1k tokens on the 11 BIRD Mini-Dev databases while carrying more facts. Sources can now return polars or pyarrow frames instead of only pandas. The release requires Python 3.11+, Panel 1.9+ and Bokeh 3.9+. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I run Lumen's lumen-ai serve without setting up an API key? Lumen 1.4 supports four API-key-free ways to run: using a coding CLI you're already signed into (Claude Code, Codex, GitHub Copilot CLI, or Antigravity CLI), Kilo's free shared model tier which needs no signup, or a local Ollama model (defaulting to a 9B model via 'ollama pull qwen3.5:9b'). If nothing is configured, Lumen prints which options would work on your machine instead of guessing. _Developers testing LLM-powered data tools without burning API credits can track releases like this on daily.dev._ ### Why did my table context for LLM agents use so many tokens before, and how was it fixed in Lumen 1.4? Lumen's old table context was built from a shuffled 100-row sample lacking SQL types or keys, which could list wrong ranges, drop NULLs, and change on every run, defeating prompt caching. Lumen 1.4 profiles tables with the cheapest accurate strategy per source, an exact pass for tables up to a million rows, shrinking context from 22.4k to 16.1k tokens on 11 BIRD Mini-Dev databases while carrying more facts. _Anyone optimizing LLM token costs for data agents can follow framework changes like this on daily.dev._ ### Can Lumen use polars or pyarrow dataframes instead of pandas? Yes, as of Lumen 1.4, any dataframe library supported by narwhals is accepted, including polars and pyarrow, and schema generation, Filter, and common transforms run on them natively without converting to pandas first. Some operations still convert to pandas, such as Aggregate unless with_index=False, and anything touching the pandas index; views also convert since hvPlot, Tabulator, and Vega only read pandas. _Developers choosing between polars and pandas for data pipelines can track library support changes on daily.dev._ ## Similar posts on daily.dev - [What's New with Claude Opus 4.6](https://daily.dev/posts/what-s-new-with-claude-opus-4-6-wzhueggzn) · DigitalOcean Community · 0 upvotes · 0 comments - [Untitled](https://daily.dev/posts/untitled-sn7w6mzum) · SitePoint · 0 upvotes · 0 comments --- Tags: [#python](https://daily.dev/tags/python), [#ai-agents](https://daily.dev/tags/ai-agents), [#data-visualization](https://daily.dev/tags/data-visualization), [#polars](https://daily.dev/tags/polars) [View this post on daily.dev](https://daily.dev/posts/lumen-1-4-0-release-oivtn6yi5) ```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":"Lumen 1.4.0 Release","url":"https://daily.dev/posts/lumen-1-4-0-release-oivtn6yi5","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/lumen-1-4-0-release-oivtn6yi5"},"datePublished":"2026-10-26T19:53:37.009Z","dateModified":"2026-10-26T19:57:21.692Z","description":"Lumen 1.4 makes the conversational data analysis framework easier to start and cheaper to run. You can now use it without an API key via Claude Code, Codex,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/a1da4730b69c0f265a333056cdc367c3?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/a1da4730b69c0f265a333056cdc367c3?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Planet Python","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":"Planet Python","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/eac3cc584cb043b5a242d13752bd3d0c","url":"https://daily.dev/sources/planetpython"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/lumen-1-4-0-release-oivtn6yi5","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"python,ai-agents,data-visualization,polars","timeRequired":"PT18M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Planet Python","item":"https://daily.dev/sources/planetpython"},{"@type":"ListItem","position":3,"name":"Lumen 1.4.0 Release"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/lumen-1-4-0-release-oivtn6yi5#faq","mainEntity":[{"@type":"Question","name":"How can I run Lumen's lumen-ai serve without setting up an API key?","acceptedAnswer":{"@type":"Answer","text":"Lumen 1.4 supports four API-key-free ways to run: using a coding CLI you're already signed into (Claude Code, Codex, GitHub Copilot CLI, or Antigravity CLI), Kilo's free shared model tier which needs no signup, or a local Ollama model (defaulting to a 9B model via 'ollama pull qwen3.5:9b'). If nothing is configured, Lumen prints which options would work on your machine instead of guessing. Developers testing LLM-powered data tools without burning API credits can track releases like this on daily.dev."}},{"@type":"Question","name":"Why did my table context for LLM agents use so many tokens before, and how was it fixed in Lumen 1.4?","acceptedAnswer":{"@type":"Answer","text":"Lumen's old table context was built from a shuffled 100-row sample lacking SQL types or keys, which could list wrong ranges, drop NULLs, and change on every run, defeating prompt caching. Lumen 1.4 profiles tables with the cheapest accurate strategy per source, an exact pass for tables up to a million rows, shrinking context from 22.4k to 16.1k tokens on 11 BIRD Mini-Dev databases while carrying more facts. Anyone optimizing LLM token costs for data agents can follow framework changes like this on daily.dev."}},{"@type":"Question","name":"Can Lumen use polars or pyarrow dataframes instead of pandas?","acceptedAnswer":{"@type":"Answer","text":"Yes, as of Lumen 1.4, any dataframe library supported by narwhals is accepted, including polars and pyarrow, and schema generation, Filter, and common transforms run on them natively without converting to pandas first. Some operations still convert to pandas, such as Aggregate unless with_index=False, and anything touching the pandas index; views also convert since hvPlot, Tabulator, and Vega only read pandas. Developers choosing between polars and pandas for data pipelines can track library support changes on daily.dev."}}]} ```