Skip to content
zeekk0Public

About

OpenAmika: MekaHime's AI companion Amika, open source. Live2D avatar, real-time voice, long-term memory, OpenRouter or Ollama brain.

Topics

Resources

Stars

666 stars

Watchers

5 watching

Forks

Latest commit

 

History

76 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Amika

OpenAmika

Hi! I'm Amika — MekaHime's AI companion, open source

简体中文 | 日本語 | English | Tiếng Việt | 한국어

Made with Love GitHub Release License: MIT Platform Python 3.11 | 3.12

Hugging Face Discord Website Try Online Tests GitHub stars

Install her: 7 steps

Before you start: Windows 10 or 11, an NVIDIA graphics card (RTX 20-series or newer, 8 GB of video memory) and about 40 GB of free disk?

Rather watch it? The same 7 steps as a 3-minute video, with fixes and answers for when you get stuck: mekahime.com/openamikatutorial.

STEP 1 — Install Python 3.11 and Git

Takes about 5 minutes.

  1. Click python-3.11.9-amd64.exe. Open it. Tick Add python.exe to PATH (bottom of the first window). Click Install Now.
  2. Click Git-2.56.0.4-64-bit.exe. Open it. Click Next until it says Install, click Install, then Finish.

You see: "Setup was successful" (Python) and "Completing the Git Setup Wizard" (Git).

STEP 2 — Open PowerShell in the folder where she will live

  1. Press the Windows key, type powershell, press Enter. A normal window: not "Run as administrator".
  2. Copy the line below. Click into the blue window. Right-click (that pastes). Press Enter.
cd C:\

You see: PS C:\> and a blinking cursor.

Any drive works: cd D:\ puts her on D:. Short paths work best. Not inside OneDrive.

STEP 3 — Download OpenAmika

Paste each line, press Enter, wait for PS ...> to come back. Takes about a minute.

git clone https://github.com/zeekk0/OpenAmika.git
cd OpenAmika

You see: PS C:\OpenAmika>.

  • If you see 'git' is not recognized → close this window, do STEP 2 again (a new window finds Git), then this step.
  • If you see destination path 'OpenAmika' already exists → she is downloaded already: paste cd OpenAmika and go on.

STEP 4 — Run the setup

powershell -ExecutionPolicy Bypass -File scripts\setup_windows.ps1

Takes 15 to 60 minutes: about 11 GB of downloads. Leave the window open.

First it asks for a Hugging Face token: press Enter to skip, or paste a free Read token from huggingface.co/settings/tokens (it makes the downloads much faster).

You see at the end: OpenAmika setup summary, lines starting with OK, then Done. Next:.

  • If you see The argument 'scripts\setup_windows.ps1' to the -File parameter does not exist → this window is not in her folder. Paste cd C:\OpenAmika (or the cd line her setup printed), press Enter, run the line again.
  • If you see Python 3.11 was not found → STEP 1 again, with Add python.exe to PATH ticked. Then a new window (STEP 2), cd C:\OpenAmika, and this line again.
  • If the download crawls (about 50 KB/s) → press Ctrl+C, run the line again and paste a Hugging Face token when it asks. It goes on where it stopped.
  • If a line starts with !! → do what that line says, then run the line again. Finished parts are skipped.

STEP 5 — Her voice and her model

envs\brain\Scripts\python.exe scripts\setup_tts.py

Takes 10 to 30 minutes: about 5.5 GB of downloads. It shows her license, then asks: type yes, press Enter.

You see: OK voice MekaHime_Amika and OK Live2D model Amika, then Done. Next:.

  • If you see not downloaded: her license was not accepted → run the line again and type yes.
  • If you see 'envs\brain\Scripts\python.exe' is not recognized → this window is not in her folder (cd C:\OpenAmika), or STEP 4 did not finish.

STEP 6 — Start her

.\Start_OpenAmika.bat

Next time, just double-click Start_OpenAmika.bat in C:\OpenAmika.

Takes 1 to 3 minutes (the first start is the slowest).

You see: a window called OpenAmika with green rows. Your browser opens her page (http://localhost:53000) and her first-run window asks your name and your city.

  • If her window says moved to … in settings.json (chat_port) → Windows had reserved port 53000, so she took the next free one. Open the address at the top of her window.
  • If her window stops on an error → it stays open and names its log (logs\launcher.log). Bring that file to the Discord.

STEP 7 — Paste your OpenRouter key in her first-run window

  1. On openrouter.ai: sign in, open Credits, add a few dollars. OpenRouter is prepaid: a key alone is not enough.
  2. Open Keys → Create Key. Copy the key (it starts with sk-or-v1-).
  3. In her first-run window, at her brain: click OpenRouter, paste the key, click Save key.

You see: a green ✓ Key saved … · $5.00 of credit left. Click Next. Click her page once (so the browser lets her speak), press the power button, and talk.

  • If you see no credits → add credits on openrouter.ai, then click Save key again.
  • If a reply fails → the chat shows OpenRouter's own reason. A model that is gone: Setup → Brain → Switch back to the OpenRouter preset.

Downloaded the ZIP instead of git clone? Unzip it. Open the inner folder (the one with Start_OpenAmika.bat in it). Right-click an empty spot → Open in Terminal. Then do STEP 4 to STEP 7. (Windows 10: click the folder's address bar, type powershell, press Enter.)

STEP 1 to STEP 6 follow the step-by-step guide that metro01 wrote for the MekaHime Discord's help desk. Thank you, metro01!

Updating her (keeps her memory)

Her memory is in memory_f and brain\memories. Back them up before anything.

I used git clone

  1. STEP 1 — Stop her. Double-click Stop_OpenAmika.bat.
  2. STEP 2 — Double-click Update_OpenAmika.bat in her folder. It saves her memory and your settings to backups\update-<date>.zip, downloads the update and installs what changed. Type yes if her license is asked.

You see: Updated. Start her: double-click Start_OpenAmika.bat.

  • If you see She is running → STEP 1 first.
  • If it lists files "with your own changes" → your version is kept. Her memory is safe.

No Update_OpenAmika.bat in her folder? You installed her before v1.1.0 (9 November 2026). Do STEP 2 this way, one time. Open her folder, right-click an empty spot → Open in Terminal (Windows 10: click the folder's address bar, type powershell, press Enter). Paste each line and press Enter:

git fetch origin
git show origin/main:scripts/update_openamika.ps1 > scripts\update_once.ps1
powershell -ExecutionPolicy Bypass -File scripts\update_once.ps1

It is the same update, with the same backup. After it, Update_OpenAmika.bat is in her folder.

I downloaded the ZIP

  1. STEP 1 — Stop her. Double-click Stop_OpenAmika.bat.
  2. STEP 2 — Download and unzip the new ZIP next to the old folder.
  3. STEP 3 — Copy these from the old folder into the new one (the ones you have): memory_f, brain\memories, settings.json, brain\.env, config\seed.json, config\ollama.json, config\voice_map.json, config\tools.json, config\messengers.json.
  4. STEP 4 — Run the setup in the new folder: STEP 4 and STEP 5 of the install, from the new folder. It is faster the second time.
  5. STEP 5 — Start her from the new folder. She may ask your name and city again; her memories are all there. Delete the old folder once she remembers you.

About her

Amika is a Live2D anime girl who talks with you by voice and text in real time, remembers you, and lives with you in your city. She runs on your own PC, with OpenRouter or fully local with Ollama. A hosted version runs at mekahime.com/amika.

💖 If Amika makes your day better, give us a ⭐ on GitHub — it helps more people find her! · Support the project: https://www.mekahime.com/donate

What you need

  • Windows 10 or 11 and an NVIDIA graphics card, RTX 20-series or newer, with a recent driver. Video memory (VRAM) is what counts, not your PC's RAM:

    how she runs her own parts her language model for example
    Semi-Local (recommended) about 5.5 GB of video memory: her ears, memory and voice. An 8 GB card on OpenRouter: none on your card an RTX 3060 laptop runs her: her voice comes a few seconds after the text, and a user's bill read "only 2 cents so far"
    + her camera eyes about 4.3 GB more: a 16 GB card
    Fully Local (Ollama) 5.5 GB (about 10 GB with her eyes) the rest: about 19 GB more for a 27B model one 32 GB card (RTX 5090), or Ollama on a second PC
  • Python 3.11 (3.12 works too; 3.13 and newer do not, because her voice server needs numpy 1.26.4, which has no wheels for them) and Git: STEP 1.

  • About 40 GB of free disk: a full setup takes about 37 GB (about 28 GB without her camera eyes), and pip cache purge gives back about 6 GB of it afterwards. It downloads about 16 GB and takes 30 to 60 minutes.

  • An OpenRouter key with a few dollars of credit for her language model, or Ollama with a model of 27B or more and a graphics card big enough for it (see Which language model?).

Which language model?

Her ears, voice, memory and eyes always run on your PC. Her language model, the part that thinks and writes her replies, runs one of two ways:

  • Semi-Local (recommended): through OpenRouter on your own key. In almost all cases this is the better choice: she uses the models she was tested with, answers fast, and it only costs a few dollars a month (about $1.50 if you talk 30 turns a day, about $5 at 100 turns a day).
  • Fully Local: on your PC with Ollama. Use a model of 27B parameters or more; 70B or more gives the full, reliable experience. Models smaller than 27B may be really bad: she loses her character, forgets what you told her and breaks her reply format (her first-run window and Setup → Brain warn you). A local model is also slower: all her thinking takes turns on it, so after you interrupt her, or while she saves a memory or reflects, a reply can take 5–50 s.

What OpenRouter is, and what it costs. OpenRouter is one account for many AI models: with your key, she sends her thinking to the models she was tested with, and OpenRouter charges their prices. It is prepaid: add a few dollars of credit first. A key on an account with no credits makes every reply fail (her chat then says so). One exchange costs about $0.002 (MekaHime's own ledger: $0.00165 a turn on the default models): 30 a day is about $1.50 a month, an hour of talking every day about $5. A long day can reach 50 cents: an hour of voice chat is about 100 turns, every turn sends her whole character and the memories it needs again (about 8,000 tokens), she reflects and writes memories between turns, and a helper (Hermes, nanoMuse) runs its own model on the same key. Your own figure is on the Use page, in the Session cost tile.

A local model needs its own video memory on top of the ~5.5 GB her ears, memory and voice use (~10 GB with her eyes): about 19 GB for a 27B model such as gemma3:27b, 24 GB for a 32B one such as qwen3:32b, and 50 GB for a 70B one such as qwen2.5:72b. So one 32 GB card (RTX 5090) holds a 27B model next to all of her (a 32B one only with her eyes off); a 24 GB card does not. A 70B model fits next to her only on a 72–96 GB workstation card. Otherwise run Ollama on a second PC (the Setup page takes its address), or stay Semi-Local. Whichever model you run, pick one that supports Chinese, since her prompts are written in it (llama3.3:70b, for one, does not). How to switch: Optional add-ons.

Local models tried with her (Ollama sizes as ollama.com lists them; "loaded" is with her 16k context):

model Ollama download video memory, loaded how it went
Gemma 4 26B MoE (26B, about 4B active per token) gemma4:26b 16–19 GB about 21 GB holds her character; the one model under 27B that does. A mixture of experts: a user ran it with 6–10 GB of video memory and the rest on the CPU
Qwen 2.5 72B qwen2.5:72b 47 GB about 50 GB the full experience: a 72–96 GB card, or a second PC
GLM 5 (744B, 40B active) none as a local model (Ollama retired glm-5 on 2026-07-15; its GLM 5.x tags are cloud ones) – about 450 GB at 4-bit: a multi-GPU server her default deep model on OpenRouter, where it costs a few dollars a month

What she can see and control

what what she does with it the switch
Your microphone hears you while her power is on; speech recognition runs on your PC power and mic buttons on the Use page
Your webcam (her eyes) sees you only while you turn the camera on; a small model on your PC describes the picture, and only that description reaches her language model the camera button on the Use page; "vision": {"enabled": false} in settings.json removes her eyes
Your screen and your PC (browser, files, email drafts) only through a helper, Hermes Agent or nanoMuse, when you ask her for an errand; risky steps wait for your OK Tools page: the helper's switch (off until you set one up)
Your smart home (Zigbee plugs) switches only plugs you paired, only after you accepted the IoT safety statement Tools page: the IoT switch; IoT\Setup_IoT_admin.bat is optional, only for this
Discord and Telegram text only, through your own bot on your own server; only user ids you allow reach her Setup → Text her

Semi-Local, the text of what you say, her memories that the turn needs and her camera's description go to OpenRouter; your voice and the camera pictures stay on your PC. Fully Local (and without a helper), nothing leaves your PC.

To start her memory over later: Setup page → Danger zone → Wipe memory. She forgets every conversation and fact, and a backup stays in memory_f\backups\. If she is still talking, it waits a few seconds, then offers Force wipe (she stops what she is doing). From a script: envs\brain\Scripts\python.exe launch.py --wipe-memory (--force to stop her first; it also works while she is not running).

Talk to her

Double-click Start_OpenAmika.bat in the OpenAmika folder (in PowerShell: .\Start_OpenAmika.bat). One console starts all of her, and your browser opens http://localhost:53000 when she is ready (if Windows reserved that port, she moves to the next free one and her console shows the address): about a minute, a few more the very first time. On your first visit she introduces herself and asks your name, your city and, if brain\.env has none yet, your OpenRouter key. Then press the power button and talk, or type. Click the page once so the browser lets her speak. When you go quiet she may say something, counted from the moment her voice stops; Setup → You & Amika → "When you go quiet" sets it to Normal, Rare or Off.

The light at the top right of every page is her health: green means everything works; click it to see what is not. Double-click Stop_OpenAmika.bat (or press Ctrl+C in her console) to stop her. To check everything without starting anything, run envs\brain\Scripts\python.exe launch.py --check: it ends with result: ready, or names what to fix.

To update her: Updating her.

If something goes wrong

what you see what to do
The argument 'scripts\setup_windows.ps1' to the -File parameter does not exist Your window is not in her folder: cd C:\OpenAmika (a ZIP: the inner OpenAmika-main folder), then the line again.
Python 3.11 was not found Install Python 3.11 (tick "Add python.exe to PATH"; 3.12 works too), open a new PowerShell window and run the setup again.
The downloads crawl (about 50 KB/s) Hugging Face throttles downloads without a token. Ctrl+C, run the setup again, paste a free Read token from huggingface.co/settings/tokens when it asks: it continues where it stopped.
!! voice MekaHime_Amika not downloaded: her license was not accepted Run envs\brain\Scripts\python.exe scripts\setup_tts.py again and type yes.
!! voice MekaHime_Amika Hugging Face refused the download (HTTP 401) Her files are not public yet, or your HF_TOKEN is wrong. Set a read token from huggingface.co/settings/tokens in the same window ($env:HF_TOKEN = "hf_...") and run the command again.
download failed after 4 tries or Check the internet connection Run the same command again: it continues where it stopped. If huggingface.co cannot be reached from your network, set $env:HF_ENDPOINT to a Hugging Face mirror first; every file is still checked.
does not see the GPU (CUDA) Update your NVIDIA driver and run the setup again. Without an NVIDIA card: powershell -ExecutionPolicy Bypass -File scripts\setup_windows.ps1 --cpu.
The filename or extension is too long The folder's path is too long for Windows: move OpenAmika to a short folder such as C:\OpenAmika and run the setup there.
settings.json not found Run powershell -ExecutionPolicy Bypass -File scripts\setup_windows.ps1 first.
No Live2D model yet on the stage, or !! Live2D model / !! her voice in launch.py --check Run envs\brain\Scripts\python.exe scripts\setup_tts.py.
no OpenRouter key yet (launch.py --check, the setup summary) Paste your key in the first-run window, or on Setup → Brain.
Her console says moved to … in settings.json, or [Errno 13] … 53000 Windows reserved that port (Hyper-V, WSL2, Docker Desktop). She moves to the next free port by herself: open the address at the top of her console.
!! TTS (voice) port 8002 answers but has no voice MekaHime_Amika Another voice server uses her port (OpenSakura's uses 8002 too). Stop it, or give her another port: change 8002 in tts_url and in services.tts (--port, health) in settings.json.
Every reply fails The chat shows the provider's reason: no credits (OpenRouter is prepaid: add a few dollars at openrouter.ai/settings/credits), a model OpenRouter no longer offers (Setup → Brain → "Switch back to the OpenRouter preset"), a wrong key (paste a new one on Setup → Brain: it works at once, no restart).
She answers in text but you hear nothing Give her voice about 20 seconds after a start, then click the page once. The health light must show TTS green.
Her voice spells a word in CAPITALS letter by letter Fixed: she lower-cases shouting before she speaks. An acronym she should spell goes into "tts_keep_caps" in config/word_replace.json.
Her voice stutters or is very slow Close other programs that use the graphics card, or turn her eyes off: "vision": {"enabled": false} in settings.json.
On Ollama she is out of character, forgets things, or her replies come out broken The model is too small. Use one of 27B or more (70B or more for the full experience), or go back to OpenRouter: Setup → Brain → OpenRouter → "Switch back to the OpenRouter preset".
On Ollama she answers very slowly If ollama ps shows a CPU/GPU split instead of 100% GPU, the model does not fit in your video memory next to her. With 100% GPU, slow replies after you interrupt her, or while she saves a memory or reflects, are normal for a local model. See Which language model?.
Her console closes before you can read it It stays open now after an error and names logs\launcher.log.
Amika is busy when you wipe her memory or reset everything It waits a few seconds for her to finish; if she is still busy, press Force wipe (she stops what she is doing). The backup is kept either way.
Model selection is busy on the Setup page Press Unlock next to the message: she stops what she is doing, and the switch goes through.
The console's brain row (or the Setup page's brain start) is red The row says why: the port is taken, the environment is broken, the model does not fit, or the key is missing.
A helper's card on the Tools page is red, or she says she cannot reach Hermes or nanoMuse Run envs\brain\Scripts\python.exe launch.py --check and look at its agent and hermes key lines; every message and its fix is in docs/TOOLS.md.

Every part writes a log in logs\. For anything else, run envs\brain\Scripts\python.exe launch.py --diagnose: it writes one zip in logs\ (the startup checks, the end of each log, your settings without keys, versions, free disk and GPU memory). Attach it when you ask on Discord or open an issue. Using an AI coding agent? Point it at README_IF_YOU_ARE_AI_AGENT.md.

Optional add-ons

  • Fully Local: install Ollama, run ollama pull gemma3:27b (17 GB; or another model of 27B or more, see Which language model?), then on the Setup page choose Brain → Ollama → Detect models → Use this model. No key needed. For Ollama on another PC, type its address there before Detect models.
  • Her helpers: Hermes Agent or nanoMuse run errands for her, and her desk view shows what they do (below). Any other MCP server works too, on the Tools page (docs/TOOLS.md).
  • Smart home (optional: nothing else needs it): Zigbee plugs she can switch (IoT/README.md). IoT\Setup_IoT_admin.bat installs what it needs as administrator; skip it if you have no Zigbee plugs. It stays off until you read and accept the IoT safety statement (docs/IOT_DISCLAIMER.md) on the Tools page. She runs on language models and can make mistakes: never connect heaters, locks, valves or anything that could cause harm.
  • Text her on Discord or Telegram (docs/MESSENGERS.md). Her bot lives on your own Discord server: the bot must share a server with you before a DM works.
  • Your phone: Tailscale_Serve_OpenAmika.bat makes her reachable on your own devices over Tailscale.
  • Desktop pet: she stands on your desktop by herself, always on top of your other windows. Press Desktop pet on the Use page (below).

Desktop pet mode

She can also stand on your desktop by herself: a see-through window with only her in it, on top of your other programs, with her voice, lip-sync and expressions as on the Use page. Windows only.

From the Use page (the easy way): open the expand menu at the top-right of her stage and choose Desktop pet.

  1. The first time, the page asks for one small install (pywebview, a few megabytes). Press Install and open.
  2. Her window opens on your desktop and the conversation goes with her: if she was on, she is on there; if your microphone was on, it is on there.
  3. The page now says Amika is on your desktop. Press Bring her back here and she is on the page again.

The button shows only in a browser on the PC that runs her (her window opens on that PC's desktop), so not on your phone.

If Voice output is off, she asks before moving: Enable and move turns spoken replies on for the pet; Cancel leaves the browser session untouched. A failed move keeps the previous voice setting. After a successful move, spoken replies stay enabled when you bring her back to the browser.

The microphone and audio output selected in the web page carry over to her desktop window. Hover the microphone for its input name, or power for the output name. A selected device that the page cannot find (unplugged, or the browser has not been allowed to read device names yet) moves as System default, and the page tells you; the selection itself stays saved. If her desktop window cannot find a named device, or finds it twice, the move fails and the browser session is restored. System default stays System default. The webcam switches off after a successful move; bring her back and enable it again when you want to use it. Pet controls keep their softer styling.

Straight to the desktop: double-click Start_OpenAmika_Pet.bat instead of Start_OpenAmika.bat (or run launch.py --pet): when she is ready, her window opens instead of the browser page. This way does not install anything by itself. Without the install the page opens in your browser as usual, and its Desktop pet button does the install; or do it by hand, once:

envs\brain\Scripts\python.exe -m pip install -r brain\requirements-pet.txt

Hover over her for a small bar: power, microphone, the full page, close. Click the microphone and talk as usual. Drag her to move the window; the mouse wheel over her makes her bigger or smaller; she comes back where you left her. Closing her window does not stop her: the Desktop pet button (or envs\brain\Scripts\python.exe pet.py) opens it again, Stop_OpenAmika.bat stops everything.

  • One conversation at a time: while she is on your desktop, the page waits for her. Bring her back here on the page, or the full-page button on her window, moves her back.
  • Her desktop window has no text box: talk to her, or bring her back to the page to type.
  • Meet her on the full page first: the first-run wizard, the camera, the background photo and her helpers' paper doll are there.
  • The see-through corners of her window still take your clicks: keep it snug around her.

Her helpers: Hermes and nanoMuse

She can hand errands to a helper program on your PC, Hermes Agent or nanoMuse: it searches the web, uses a browser and works with files. She decides whether to help, gives the helper the task after she has answered you, relays its questions, and tells you the result in her own words. While a helper is switched on, her desk view (what she does right now) sits at the bottom right of her stage: MekaHime's little folded-paper doll acts out the helper's side, so you can see what is being done for you. Each helper runs its own model on OpenRouter, and the key she uses works for it.

Hermes Agent: install it from a PowerShell window. This form of its installer asks no questions; the lines after it do the setup OpenAmika needs:

& ([scriptblock]::Create((irm https://hermes-agent.nousresearch.com/install.ps1))) -NonInteractive

Then, in a new PowerShell window:

hermes model                                    # choose OpenRouter, paste your key, then "Enter custom model name": z-ai/glm-5
$key = [guid]::NewGuid().ToString('N')          # a random key for its API server
hermes config set API_SERVER_ENABLED true
hermes config set API_SERVER_KEY $key
hermes gateway stop                             # a Hermes already running reads the key only when it starts again

nanoMuse: clone it into the Nanomuse folder, give it a config.toml and put your key in its own vault; the commands and the file are in docs/TOOLS.md.

Connect them. In the OpenAmika folder, run Copy-Item config\tools.example.json config\tools.json (only if there is no config\tools.json yet) and notepad config\tools.json. Set "enabled": true at the top and in the helper's entry. For nanoMuse, "NANOMUSE_LANGUAGE" is the language it answers in ("en"; or "ja", "zh-CN"). For Hermes, copy its key with this line and paste it (Ctrl+V) as "HERMES_API_KEY":

(Select-String -Path "$env:LOCALAPPDATA\hermes\.env" -Pattern '^API_SERVER_KEY=(.+)').Matches[0].Groups[1].Value | Set-Clipboard

Save and start her (.\Start_OpenAmika.bat): she starts her helpers herself. On the Tools page the helper's card turns green (agent online).

Her desk view. Open the Use page: the little paper doll slides onto her stage. Ask her, for example, "Can you have Hermes check tomorrow's weather?". She answers, types the task in her speech bubble, and the desk view takes over: the doll thinks while the helper works, speaks the helper's steps in its own bubble, looks surprised when the helper has a question or needs your OK (say yes or no, or click Approve on the Use page), and is happy when the result comes, which she then tells you. It comes with setup_tts.py (OK paper doll Host in its summary).

Every step in detail, and what each message means: docs/TOOLS.md.

Licenses

  • OpenAmika's code is MIT (LICENSE). Third-party parts keep their own licenses (THIRD_PARTY_NOTICES.md); the Live2D Cubism Core in ui/vendor/ is not MIT.
  • Amika herself is not MIT. Her character, art, voice, Live2D model and their weights, and the paper doll of her desk view (MekaHime/MekaHime_Amika), belong to MekaHime Inc. and are free under a personal-use license: personal use only, 18+, no redistribution or re-hosting, and no training other models on them. Two creator rules: earn on social media (videos, streams, ads, memberships, sponsorships) with no limit, crediting MekaHime.AI visibly; and fan merch up to US$1,000 in total, as an individual, made from fan art you recreated yourself (never an original image MekaHime published on its website or social media), with "MekaHime.AI" on the visible side and a small "Fan Art" label. Any other commercial use needs a license from MekaHime (hi@mekahime.com). The LICENSE on Hugging Face governs.
  • Hermes Agent (MIT) and nanoMuse (GPL-3.0-or-later) are separate programs you install yourself; OpenAmika only talks to them.
  • Built on SenseVoice, GPT-SoVITS, bge-m3, Qwen3-VL, pixi-live2d-display with the Live2D Cubism SDK, OpenRouter, Ollama and the Model Context Protocol. Thank you.

Support

Questions and bugs: GitHub issues, hi@mekahime.com, or Discord. Support the project: https://www.mekahime.com/donate

MekaHime

About

OpenAmika: MekaHime's AI companion Amika, open source. Live2D avatar, real-time voice, long-term memory, OpenRouter or Ollama brain.

Topics

Resources

Stars

666 stars

Watchers

5 watching

Forks

Releases

Packages

Contributors

Languages

⌂