Skip to content

Releases

EZLM 0.1.12 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 06 Nov 22:04

Connect now sets up OpenCode, Hermes and DeepSeek Harness for you, checks that each one really works with your model, and opens them in one click.
Updating keeps your models, keys, settings and chats. Press Update in Settings.

This release is for Windows (regular and ARM PCs). The Mac app stays on 0.1.11 for now; a Mac build of these changes follows.

What's new in 0.1.12

  • Connect, rebuilt. Pick an app, press Connect, and EZLM:
    • installs the app if you don't have it, at a tested version, and checks the download before using it;
    • adds your model to the app as "EZLM", without touching your other settings;
    • asks the app to read a file with one of its tools and answer, so "Connected" means it really works;
    • shows each step as it goes, and a plain reason if something fails, with Show details and Report a problem.
  • DeepSeek Harness. A third app on Connect. Open it in your browser, or with Open in terminal in a terminal app (dsh-TUI), which EZLM installs the first time.
  • Open buttons on Overview. Open OpenCode, Hermes or DeepSeek Harness straight from Overview. EZLM starts your model first if needed, and tells you if an app closes right after opening, and why.
  • OpenCode:
    • EZLM uses your own OpenCode when it works with your chat history, and otherwise installs its own OpenCode 2.2.5 (an older OpenCode can't open the history a newer one wrote).
    • A first answer that took minutes on small thinking models now comes in seconds (chat titles no longer make the model think).
  • Hermes:
    • Works with Hermes installed from its script or from the Microsoft Store, and never installs Hermes's background service or a login item.
    • If Hermes needs a larger context than your model has, the tile says so and Connect again makes the room.
    • The first check on a new Hermes can take a few minutes while Hermes sets up its tools; Connect now says so.
  • Memory. EZLM now counts the memory your model's conversations keep between turns, and gives the rest to llama-server's prompt cache instead of a fixed 8 GB. A full cache can no longer push your computer into swap. On a few computers this picks a different model than before (for example, a 12 GB graphics card now gets a model that fits on the card instead of one that pages from disk).
  • Clearer progress. Installer warnings that don't mean anything is wrong stay in the log instead of showing as progress, and the first-answer estimate is closer for waits around a minute.
  • Windows on Arm. OpenCode, Node.js and DeepSeek Harness install as native Arm versions.
  • Overview lists the apps you've connected, and Settings > About credits the apps' makers.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): this release has no Mac build. Use EZLM 0.1.11 from https://github.com/picom-code/ezlm/releases/tag/v0.1.11-beta for now.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, on the error itself, or on an app's tile). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe ED9B10A2B80120A9049D2CA6C7686ECE26F6370258DEBE3B2432C5490E230FD0

EZLM 0.1.11 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 04 Nov 22:38

Chat comes to the Mac, and Windows PCs with AMD or Intel graphics now use their graphics card.
Updating keeps your models, keys, settings and chats. Press Update in Settings.

What's new in 0.1.11

  • Chat on the Mac. Everything the Windows app's Chat page does, now on the Mac too:
    • talk to your model right inside EZLM, with no keys to copy;
    • the live signal across the top shows what your model is doing, even when another app is using it;
    • answers stream in with proper formatting: lists, tables, and code panels with a Copy button;
    • thinking models show their thinking as it happens, then fold it away, and a ∴ thinking switch turns it off;
    • it remembers you: tell it about yourself, or use /remember, and open ✦ memory to see or edit everything, add your own instructions, or turn it off;
    • chats are kept (History), and you can attach text files (drag them in or press + file), Retry, Listen to an answer, switch models (Model), and type / for commands.
  • Windows with AMD or Intel graphics. PCs with an AMD Radeon RX or Pro card, an Intel Arc card, or Ryzen AI Max graphics now run models on the graphics card (through Vulkan), where before EZLM used only the processor. You can watch its memory fill as the model loads. If the card's driver can't load a model, EZLM runs it on the processor instead and says so on Overview, with a button to try the card again (after a driver update, say). Everyday built-in graphics still use the processor, which is about as fast there.
  • Chat, on both computers:
    • Long chats keep going. When a chat grows longer than your model can read at once, the oldest messages are left out and the chat carries on, instead of asking you to start over. Answers that needed this say "older messages left out".
    • Continue. An answer that runs into the model's limit says so, and Continue picks it up from where it stopped (it's there after you press Stop, too).
    • Answer without thinking. If a thinking model is still thinking after 40 seconds, or thinks until it runs out of room, one click gets you a straight answer.
    • Sharper memory. It picks up your name, never saves a "none" by mistake, drops stray formatting, and /forget age no longer takes "manager" with it.
    • ↑ in an empty message box brings back your last message with its attached files, to fix and send again.
    • Retry keeps your question and its files; an answer that's still coming in is saved if you open another chat, start a new one, or close EZLM; double-clicking Send no longer stops the answer it just started.
    • Models that don't fit this computer are greyed out in the Model menu, text files in other encodings (UTF-16, older Windows ones) attach as readable text, and with no model set up yet, Chat has an Open Models button.
  • Every model card shows its context, for example "64k context, works with coding agents", so you can see how much each model reads at once before you pick it.
  • Steadier. A model that stops while loading is noticed in seconds (before, a crash with no message meant waiting up to 10 minutes). If EZLM crashed, the next Report a problem (with details on) says what went wrong and where, without your files, keys or anything personal.
  • Windows:
    • In Chat, start typing after clicking a button (Model, History, memory) and the words go into the message box.
    • If a small model leaks its internal "think" tags into an answer, Chat hides them.
    • On Windows set to German, French, Japanese or any other language, System and the model picker now show which apps are using your graphics card.
  • Mac:
    • Opened EZLM straight from the download without dragging it into Applications? It now says so, since it would be gone after a restart.
    • Opening an old chat can't crash EZLM, big files attach without freezing it, scrolling up to read stays put while an answer streams, the newest reply stays in view after a very long answer, and Chat uses less battery.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe F88553080315AFBD4E054DAE30F3020001A8B2B4B5544D78329B81908CD1B6BE
  • EZLM.dmg 11A29C9C553B9FFA0F025930767DED6A935B26CD73A9F015F91456A11FD56BB8

EZLM 0.1.10 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 20:41

A fix for Windows user names with accents or non-Latin letters.
Updating keeps your models, keys, settings and chats. Press Update in Settings.

What's new in 0.1.10

  • Works the same for every Windows user name. If your Windows user folder has letters like é, ü, ø, or Chinese, Japanese or Korean characters (C:\Users\José, C:\Users\Müller), the model server couldn't write its log, so EZLM couldn't explain a model that failed to start and the Logs page stayed empty. EZLM now keeps that log itself, whatever the folder is called.
  • Includes everything in 0.1.9: the thinking switch in Chat, "Answer without thinking", and the fairer Test page.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe BB2523CDE069019D56C89022E28B6DC70A56703B29478ACC87DD91D2303A9ED9
  • EZLM.dmg A3B2126962196D107DC099A46AD82047B033DD837A7356AFE9B88A97FF18F29F

EZLM 0.1.9 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 20:29

Thinking models that don't keep you waiting.
Updating keeps your models, keys, settings and chats. Press Update in Settings.

What's new in 0.1.9

  • Thinking on or off (Windows Chat). Some models think before they answer. On a slower computer that can take minutes, so Chat now has a ∴ thinking switch at the top: turn it off for direct answers.
  • "Answer without thinking." If a model is still thinking after 40 seconds, this button appears under its thoughts: it stops and asks the same question again for a straight answer.
  • A fairer Test page. "Five" now counts as the right answer to the reasoning question, and a model that thinks past the limit gets a second try with thinking off (the result says so) instead of a fail.
  • Better picks for coding. Devstral Small 2 is now offered from its 3-bit size up: in real tests its 2-bit file answered in prose instead of using tools.
  • Chat's top bar is tidier: "Open in browser" moved into History.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe C8E04C47B17302FE35F8DAF38214EFD20B16FA1BCA6ABB6A78E2D6DBC067F4B8
  • EZLM.dmg C85340CABC1DF5DC70AE1A3397B9F58F7CEFB81624304825139E2E69897B24F9

EZLM 0.1.8 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 18:49

More you can do in Chat.
Updating keeps your models, keys, settings and chats. Press Update in Settings.

What's new in 0.1.8

  • Tell it how to answer. Open ✦ memory in Chat and write your instructions ("Keep answers short", "Explain things like I'm new to them"). They're used in every new chat.
  • Switch models without leaving Chat: Model ▾ lists the models on this computer.
  • Listen: any answer can be read aloud with Windows' own voice.
  • Type / for commands: a list appears as you type, and Tab finishes the command (/remember, /forget, /memory, /new).
  • Copy a whole chat from History, to paste it anywhere.
  • When a long answer finishes while you're in another window, EZLM's taskbar button lights up.
  • Includes everything in 0.1.7: the Chat page, memory, saved chats, and downloads that look after themselves.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe 93767415FCB05EFA2EB265E1EE260354FC645EB33BB7CC0419F70DF31AC88B76
  • EZLM.dmg 1C7A48674BC29BFCDC4843CC678B1F52895FC51E93B0483973F0F2DF0EE6FB7F

EZLM 0.1.7 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 18:05

Talk to your model right inside EZLM, and long downloads that look after themselves.
Updating keeps your models, keys and settings. Press Update in Settings.

What's new in 0.1.7

  • Chat (Windows). A new Chat page: talk to your model without any other app, and without copying keys. Answers stream in with proper formatting: lists, tables, and code panels with a Copy button. Thinking models show their thinking as it happens, then fold it away. A live signal across the top shows what your model is doing, even when another app is using it.
    • It remembers you. Tell it about yourself and it keeps short notes for future chats (only on this computer). Type /remember to save something yourself, /forget to remove it, or open Memory to see and edit everything (or turn it off).
    • Your chats are kept. Pick up the last one next time, or open older ones from History.
    • Attach text files (drag them in or press + file), Retry an answer, Esc to stop, ↑ to bring back what you sent.
  • Downloads keep going. EZLM keeps your computer awake while it downloads or moves models (closing the lid still works). A connection that keeps dropping no longer makes it give up, and if the internet is really gone, it says so in plain words.
  • Closing EZLM mid-download asks first, so a stray click doesn't stop a long download. On a Mac, quitting asks, and the window's close button waits until it's done.
  • See progress without opening the window: on the taskbar button (Windows) or the Dock icon (Mac), with a nudge when it's finished.
  • Repairs itself: if an antivirus app removes the model server program, pressing Start server downloads it again.
  • Fixed: after closing "Free up memory" with Esc, confirmation windows stopped appearing until EZLM was restarted.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe 614CEE7D937604227FBF4AC322BA879A74CC7998241729542F2F4B56E43FAEE7
  • EZLM.dmg 547B6B0278D4460B1BCCDF2E6F9BABDF2E7FD0C6EFEC0C2EB3851B05369B84F0

EZLM 0.1.6 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 14:54

Updating is safer: if an update can't finish, EZLM opens again instead of staying closed.
Updating keeps your models, keys and settings. Press Update in Settings.

What's new in 0.1.6

  • Updates that can't leave you without EZLM. Before installing, EZLM stops its own background helpers (for example the start-up that pre-reads a big model) so the installer can replace every file. And if an update still can't finish, EZLM reopens the version you had, so you can try again from Settings.
  • Includes 0.1.5's fixes: downloads can't get stuck on a silent connection, and the beta's model list holds only models that real computers can run.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe 1A6B5911DAB894AC876C6958E84E044B63195D5EC1D87B9A06A7E60BE07AA34F
  • EZLM.dmg 6937F38FE707778DF513BFCD98E098A8CDEECF7454D3F09E266061CE85994639

EZLM 0.1.5 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 14:45

Downloads no longer get stuck on a silent connection, and the beta's model list is down to models that real computers can run.
Updating keeps your models, keys and settings. Press Update in Settings.

What's new in 0.1.5

  • Downloads can't get stuck. If the download server stops sending data while the connection stays open, EZLM cuts it after 45 seconds and resumes where it stopped, instead of waiting forever.
  • A tested model list for the beta. The 100 GB+ models (Kimi K3, GLM-5.3, DeepSeek V4.1 Flash, Ling 3.1 Flash, MiMo V2.6 and others) are hidden until they've been tested on hardware big enough to run them; no beta tester's computer could run them anyway. Every model that's left runs in EZLM's engine, and each is being tested with real weights; if one fails, it's removed through the online model list, with no app update needed.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). It opens a filled-in report on EZLM's public GitHub page; press Submit new issue there. Your keys and files are never included.

SHA-256

  • EZLM-Setup.exe DE8E8742E6C42A0B49CCC3CA42F60D4C65C45284941109F597CA1FF93014B99F
  • EZLM.dmg 770351E63DB1AC8A93EA74EDF633E56CCC3AE8F576CE7483ECFE143994303C54

EZLM 0.1.4 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 07:44

Every model EZLM offers has now actually run in the AI engine it ships with.
Updating keeps your models, keys and settings. Press Update in Settings.

What's new in 0.1.4

  • Every model run, not just checked. Each distinct model file EZLM can download (91 of them, from 0.5 GB to 594 GB, including Kimi K3, GLM-5.3, Ling 3.1 Flash and MiMo V2.6 Pro) was loaded and run in EZLM's engine (llama.cpp b11960). Two that couldn't load were removed: MiniMax M3.1 Flash, and three DeepSeek V4.1 Flash files (DeepSeek is still available in four other sizes).
  • The model list built into the app now matches the online one, so it's right even on networks that block GitHub.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). Reports are posted on EZLM's public GitHub page; your keys and files are never included.

SHA-256

  • EZLM-Setup.exe 32BEC761E109AF9C123E930E791A80C2BA0755969DE44CC7D90D7B0175813791
  • EZLM.dmg 5BF1303D40D479F31DF18C74862DA4A68134612CAB74D9954CDBD52C48D5B88A

EZLM 0.1.3 (beta)

Choose a tag to compare

@Picom-code Picom-code released this 03 Nov 05:48

Every model EZLM offers is now checked against the AI engine it ships, and memory planning is more accurate for Gemma 4.
Updating keeps your models, keys and settings. Press Update in Settings.

What's new in 0.1.3

  • Every model file verified. All 293 model files EZLM can download were checked against its built-in engine (llama.cpp b11960): the file exists, its format and model type are supported, and its memory needs match what EZLM plans for. Mistral Large 4 and a few files that need a newer engine were removed, so you're never offered something that can't run.
  • More accurate memory plans. Models that keep a short-term "sliding window" cache (Gemma 4, MiMo, gpt-oss and others) now have it counted, so a model EZLM says fits really fits, without slowing down or running out of memory.
  • Sturdier server start on Windows. EZLM creates its log folder if it's missing, so a start problem always comes with a reason.

Install

  • Windows 10 or 11 (regular and ARM PCs): download EZLM-Setup.exe and open it. If Windows says "Windows protected your PC", click More info, then Run anyway.
  • Mac with Apple Silicon (M1 or later): open EZLM.dmg and drag EZLM into Applications. The first time, macOS says it can't verify the app: click Done, then System Settings > Privacy & Security > Open Anyway.

Download page: https://picom-code.github.io/ezlm/

If something goes wrong

Press Report a problem in EZLM (bottom left, or on the error itself). Reports are posted on EZLM's public GitHub page; your keys and files are never included.

SHA-256

  • EZLM-Setup.exe B21512D3A023BA6DB0801F3F56D0C4E0AA722EAC44EB103BE2DABCB5501A5402
  • EZLM.dmg A752E113E63A99611B8D90CAE51BAD059666C554E71E0AD7253EE5F218FB4A48