v4.4.0

What's Changed

Turn any thread into a runnable agent.

Added

• Generate a runnable agent (Beta). Export a thread as a runnable LangGraph
  (Python) project. A step-by-step wizard scaffolds a uv project with your
  model factory, the rendered prompt (variables stay live at run time), your
  built-in tools as real code, importable stubs for custom function tools, and
  an MCP scaffold — then writes a short PLAN.md for your coding agent to finish
  and opens the project. Ships with a local web UI and debugger via
  uv run langgraph dev.

Install

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

Full changelog: https://github.com/deer-flow/llm-space/blob/v4.4.0/CHANGELOG.md

v4.3.0

What's Changed

Prompt templates grow up: Jinja2 logic, richer variables, and in-app reminders
that point you to what's new.

Added

• Jinja2 prompt templates. Prompts now support Jinja2 syntax — {% if %} /
  {% for %}, field access ({{ user.name }}), filters ({{ x | upper }}), and
  an @include("path/to/file") macro to inline another file's contents.
• New variable types. Alongside plain Text, add JSON variables
  (parsed at render into an object you can branch and iterate over; a bare
  {{ data }} prints pretty JSON) and File content variables (pick a file
  via the native OS picker or type a path; its contents are inlined at run time).

Changed

• The model selector shows its parameter buttons by default and selects the model
  name on focus for quicker editing.

Fixed

• External links are restricted to safe URL schemes.
• Root directory creation is confined to the expected location.
• Thread files are written atomically to avoid corruption on interrupted writes.

Performance

• The trace sidebar is lazy-loaded.
• The RPC stream queue is now amortized O(1).

Install

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

Full changelog: https://github.com/deer-flow/llm-space/blob/v4.3.0/CHANGELOG.md

v4.2.1

What's Changed

A small maintenance release: a new model and a couple of fixes.

Added

• Kimi K3 is now available as a built-in model.

Fixed

• Codex CLI authentication is restored.
• The tool and variable lists in the thread playground now scroll within a
  bounded height instead of pushing the layout.

Install

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

Full changelog: https://github.com/deer-flow/llm-space/blob/v4.2.1/CHANGELOG.md

v4.2.0

What's Changed

Share your threads with anyone via a link.

Added

• Share a thread. Click the share icon in the thread toolbar (top-right,
  next to run history) — or use File → Share…, or right-click a thread in the
  file tree and choose Share… — to publish it as a link anyone can open in
  their browser. The thread is published as a secret GitHub Gist under your
  account; delete the gist to revoke access.
• GitHub sign-in. Sign in to GitHub from Settings → Account. Sharing
  needs it, so the Share dialog also walks you through signing in the first time.

Install

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

Full changelog: https://github.com/deer-flow/llm-space/blob/v4.2.0/CHANGELOG.md

v4.1.1

What's Changed

A small UX release focused on the tool import dialogs.

Changed

• The built-in and MCP tool import dialogs now show per-category / per-server
  actions in the sidebar, including Enable all and Disable all for
  quickly toggling every tool in a group.
• The tools row in the thread playground now scrolls instead of pushing the
  layout when many tools are added.

Install

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

Full changelog: https://github.com/deer-flow/llm-space/blob/v4.1.1/CHANGELOG.md

v4.1.0

What's Changed

A maintenance release: no major new features, a handful of bug fixes, and three
new model/search providers.

Added

• VolcEngine Agent Plan model provider, with its models available out of the
  box.
• Brave Search as a web-search provider, selectable in Settings → Search.
• A new model for the existing Coding Plan provider.

Fixed

• Merging tabs correctly after a thread is overwritten.
• Renaming a thread no longer overwrites an existing thread with the same name.
• Deleting a prompt variable now asks for confirmation when the variable is
  referenced across multiple threads.
• Image content on assistant messages is ignored instead of causing errors.
• The active-tab ref is synced outside the render body, fixing a stale-reference
  edge case.
• Brave Search API errors are now surfaced to the user.
• Long skill descriptions no longer overflow their container.

Install

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

Full changelog: https://github.com/deer-flow/llm-space/blob/v4.1.0/CHANGELOG.md

v4.0.1

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

v0.1.1-canary.3

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

v0.1.1-canary.2

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~130 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.

v0.1.1-canary.1

Two editions. Install either, or both — they share the same ~/.llm-space data, so switching keeps your threads and settings.

• LLM Space — uses the system WebView. Small download (~27 MB), light on memory and battery.
• LLM Space Performance — embeds its own rendering engine (~127 MB). Rendering stays consistent across macOS versions, and usually performs better.

Download a DMG below. Existing installs update in place.