AmigaGPT is a versatile ChatGPT client for AmigaOS 3.x, 4.1 and MorphOS.
This powerful tool brings the capabilities of OpenAI's GPT to your Amiga
system, enabling text generation, question answering, and creative
exploration. AmigaGPT can also generate stunning images and
includes support for speech output, making it easier than ever to interact
with AI on your Amiga. Designed to integrate seamlessly with your system,
AmigaGPT delivers modern AI technology while embracing the timeless Amiga
experience.
Features:
- Generates text based on input prompts
- Generates images based on input prompts
- Shell Tool - Let the AI execute AmigaDOS commands on your system
- UI customisation
- Full conversation history
- Text-to-speech
- Custom chat server support (Google Gemini, LM Studio, Anthropic Claude,
xAI Grok)
Requirements:
- OCS/ECS/AGA Amiga or a PowerPC machine capable of running MorphO
- MUI 3 minimum; MUI 5 recommended for all features
- Motorola 68020 or higher CPU or PPC for AmigaOS 4/MorphOS
- 8MB Fast RAM minimum, 16MB or higher recommended
- For AmigaOS 3 & 4, AmiSSL 5.18 or higher
(https://aminet.net/package/util/libs/AmiSSL-v5-OS3) for OS3 and
(https://aminet.net/package/util/libs/AmiSSL-v5-OS4) for OS4
- Internet access using a TCP/IP stack such as Roadshow
(http://roadshow.apc-tcp.de/index-en.php)
- MUI 3 minimum but MUI 5 recommended for all features
MUI 5:(https://github.com/amiga-mui/muidev/releases)
or MUI 3.9:
(https://github.com/amiga-mui/muidev/releases/download/
MUI-3.9-2015R1/MUI-3.9-2015R1-os3.lha)
- **render.library 40.08** or higher
(https://aminet.net/package/dev/misc/renderlib)
- **guigfx.library 20.0** or higher
(https://aminet.net/dev/misc/guigfxlib.lha) (FPU) or
(https://aminet.net/dev/misc/guigfxlib_nofpu.lha)
- codesets.library 6.22 or higher
(http://aminet.net/package/util/libs/codesets-6.22)
- MCC_Guigfx MUI custom class for displaying images
(http://aminet.net/package/dev/mui/MCC_Guigfx)
- MCC_NList MUI custom class for lists
(http://aminet.net/package/dev/mui/MCC_NList-0.128)
- MCC_TextEditor 15.56 or higher
(https://aminet.net/dev/mui/MCC_TextEditor-15.56.lha)
- A PNG datatype - Any will do, but here is one
(https://aminet.net/util/dtype/vPNGdt.lha)
- An active OpenAI API key
- For speech on AmigaOS 3, any version of narrator.device
- For speech on AmigaOS 4, Flite Device
(http://aminet.net/package/mus/misc/flite_device)
- For OpenAI speech, AHI is required
(http://aminet.net/package/driver/audio/ahiusr_4.18)
Installation:
1. Install AmiSSL and a TCP/IP stack if not already done so
2. Download and install MUI. Version 5 recommended, version 3 minimum. Reboot.
3. Download and install codesets.library, render.library, MCC_Guigfx,
MCC_NList and MCC_TextEditor
4. Download the latest release of AmigaGPT.
5. Extract the `amigagpt.lha` archive to a temporary location.
6. Run the provided installer.
Optional steps to enable speech functionality:
AmigaGPT supports reading the output aloud. How AmigaGPT does this depends
on whether you are using AmigaOS 3 or 4. Or for OpenAI voices, this works on
every system.
Installing AHI for OpenAI voices:
If your OS does not come with AHI installed, you can get it from
https://aminet.net/package/driver/audio/ahiusr_4.18
AmigaOS 3:
Speech on AmigaOS 3 requires a file called `narrator.device` which cannot be
included with AmigaGPT because it is still under copyright. Therefore, you
must copy this file legally from yourWorkbench disks so that AmigaGPT will be
able to synthesise speech. There are 2 versions of `narrator.device`
supported, v34 and v37.
v34 is the original version that came with Workbench 1.x. v37 was an updated
version included with Workbench 2.0.x. It has more features and sounds more
natural, however it does sound quite different which is why AmigaGPT supports
you installing both versions and your choice of version to be used can be
selected in the Speech menu in the app.
Regardless of which version of `narrator.device` you choose to install
(or both), AmigaGPT requires that you install the free third party
`translator.library` v43. This works with both versions of `narrator.device`.
Installing `translator.library` v43:
Since `translator.library` v43 is not available as a standalone install, you
will need to install v42 and then patch it to v43.
* Download http://aminet.net/util/libs/translator42.lha and extract the
archive to any convenient location on your Amiga such as `RAM:`
* Navigate to that directory and double click the `Install` program
* Run the installer using all the default settings
* Download http://aminet.net/util/libs/Tran43pch.lha and once again extract it
to a location of your choice
* Navigate to that directory and double click the `Install` program
* Run the installer using all the default settings
* Reboot your Amiga - It will not work until the system is restarted
Installing `narrator.device` v34:
* Insert your Workbench 1.x disk and copy `df0:devs/narrator.device` to
`{AmigaGPTProgramDirectory}/devs/speech/34`
Installing `narrator.device` v37
* Insert your Workbench 2.0.x (you cannot use 2.1 because the speech libraries
were removed after version 2.0.4) disk and copy `df0:devs/narrator.device` to
`{AmigaGPTProgramDirectory}/devs/speech/37`
AmigaOS 4:
Speech on AmigaOS 4 is done with the help of Flite Device.
- Download Flite Device from http://aminet.net/package/mus/misc/flite_device
- Extract the archive and run the installer
Launching AmigaGPT:
* Launch the application by double-clicking the AmigaGPT icon correspondant to
the OS you are running
* You may also launch the app in the command line but before you do, run the
command `STACK 32768` to give the program 32kb of stack since the default
stack size for apps launched from the shell is 4kb and this is not enough for
AmigaGPT and will cause random crashes due to stack overflow. This is not
required when you launch the app by double clicking the icon since the stack
size is saved in the icon.
Usage:
AmigaGPT can run on its own screen or in Workbench. To configure this, open
the MUI Settings menu item from the View menu and configure it in the "Screen"
or "System" menu in the MUI Setings panel.
If you open in a new screen you have the
ability to create a screen for the app to open in. AmigaGPT supports anything
from 320x200 all the way up to 4k resolution if using a video card for RTG.
Bear in mind text will appear very tiny in resolutions above 1080p so you may
want to increase the font size settings from the MUI settings in the View menu
when the app opens.
When launching for the first time you will need to enter your OpenAI API key
before you can start chatting. If you haven't already done so, create an
OpenAI account and navigate to https://platform.openai.com/account/api-keys to
generate an API key for use with AmigaGPT.
*Chat*
When the app has opened, you are presented with a text input box. You can type
any prompt into this box and press "Send" to see the GPT model's response.
The generated text appears in the box above the input. You can choose to have
this text read aloud using the "Speech" menu option. To choose the chat model
and provider (OpenAI, Google Gemini, xAI Grok, Anthropic Claude or a custom
server), open "Chat Provider Settings" from the "Chat" menu.
To the left of the chat box is a conversation list which you can use to go to
another saved conversation. New conversations can be created with the "New
chat" button and conversations can be removed with the "Delete chat" button.
There are 2 main modes of operation: Chat and Image Generation. You can switch
between them via the tabs in the top left corner.
*Image Generation*
To generate images, open "Image Provider Settings" from the "Image" menu to
choose your provider and image model, then type your prompt in the text box
and hit the "Create Image" button. When it has been downloaded to your Amiga,
you are then able to open the image to your desired scale, or save a copy of
the file to a new location on your Amiga. Do note however that AmigaGPT will
automatically save all your generated images until you delete them. This is
just in case you would like to create a copy elsewhere.
*Shell Tool*
The Shell Tool is a powerful feature that allows ChatGPT to execute AmigaDOS
commands directly on your Amiga. This enables the AI to help you with tasks
like listing files, creating directories, managing files, and more.
To enable the Shell Tool, check "Shell Tool" in the "Chat" menu. Once enabled,
you can ask ChatGPT to perform tasks that require shell access. For example:
- "List all files in RAM:"
- "Create a new directory called Test in Work:"
- "Show me the contents of S:Startup-Sequence"
- "What version of Workbench am I running?"
For security, AmigaGPT will always show you the command the AI wants to execute
and ask for your permission before running it. You can Allow or Deny each
command.
The AI is automatically instructed about AmigaDOS command syntax, so it will
use the correct commands for AmigaOS (e.g. 'list' instead of 'ls', 'type'
instead of 'cat', etc.).
The Shell Tool also works via ARexx when using the SENDMESSAGE command with
either AmigaGPT or the AmigaGPTD daemon.
Note: Use this feature responsibly. While you have control over which commands
are executed, be cautious about allowing commands that modify or delete files.
*General*
In the "Edit" menu, you'll find basic text editing commands like Cut, Copy,
Paste and Clear. The "View" menu allows you to change the font used in the
chat and the UI.
The "Project" menu includes an "About" option, which displays information
about the program.
The "Connection" menu allows you to connect via a proxy server. It supports
both HTTP and HTTPS proxy servers but if you use an unecrypted HTTP proxy
server you can improve the performance of AmigaGPT by removing the need for
the encryption of the OpenAI traffic to be done on the system running AmigaGPT.
For an easy proxy server you can run on your local network you can try out
https://mitmproxy.org
*Chat and Image Provider Settings*
AmigaGPT supports multiple AI providers. Open "Chat Provider Settings" from
the "Chat" menu or "Image Provider Settings" from the "Image" menu.
The provider list includes built-in entries for OpenAI, Google Gemini, xAI Grok
and Anthropic Claude (chat only); these cannot be removed. For built-in
providers you can set the API key and choose the model; host, port and other
connection settings are fixed. Known models are shown for each built-in; for
OpenAI you can also use "Fetch Models" to refresh the list. You can add custom
server profiles (e.g. LM Studio) with templates and full control over all
settings. Custom profile names cannot match built-in provider names.
When you switch to another profile without saving, AmigaGPT will ask whether to
save your changes (Yes, No or Cancel).
The "ARexx" menu allows you to import and run ARexx scripts. Also, selecting
"Arexx Shell" will open a shell and put all ARexx scripts in trace mode so you
can run the scripts line by line for debugging purposes. Select "Arexx Shell"
once again to turn off trace mode and close the shell.
Note: MorphOS does not yet have a complete ARexx implementation. You must
install the 68k version of ARexx on MorphOS to be able to use ARexx with
AmigaGPT.
AmigaGPT also comes with a companion daemon app called AmigaGPTD. You may
run this in the shell. The daemon will respond to the same ARexx commands as
the main app so you do not need to launch the main AmigaGPT app.
The daemon maintains conversation history in T: so it remembers context across
multiple ARexx calls until the next reboot. This allows for follow-up questions
and multi-turn conversations. Use the NEWCHAT ARexx command to clear the
conversation history and start fresh.
Please note: AmigaGPT and AmigaGPTD listen on 2 differnt ARexx ports so
adjust your scrips accordingly. AmigaGPT listens on "AMIGAGPT" and AmigaGPTD
listens on "AMIGAGPTD".
Developing:
Source code hosted at https://github.com/sacredbanana/AmigaGPT
You can either compile the code natively or with the Docker container.
------
Native
------
*Building the AmigaOS 3 app*
If you would like to build this project from source you will need Bebbo's
amiga-gcc toolchain here https://github.com/bebbo/amiga-gcc
Once installed, get the required other SDK's (AmiSSL, Translator, json-c)
from https://github.com/sacredbanana/AmigaSDK-gcc and put these in your Amiga
dev environment created in the above step.
*Building the AmigaOS 4 app*
Get this toolchain set up https://github.com/sba1/adtools
Once installed, get the required other SDK's (AmiSSL, Translator, json-c)
from https://github.com/sacredbanana/AmigaSDK-gcc and put these in your Amiga
dev environment created in the above step.
------
Docker
------
You may use an all-in-one Docker image that is able to compile both the
AmigaOS 3 and AmigaOS 4 versions of the app.
Just install Docker on your machine and run the `build_os3.sh`,
`build_os4.sh` or `build_morphos.sh` scripts depending on which version of
the app you want to build. If you want to perform a clean build, you can set
the environment variable `CLEAN=1` for example you can run
`CLEAN=1 ./build_os3.sh`.
The build app will be saved to the `/out` directory.
Contributing:
We welcome contributions to AmigaGPT! If you have a bug to report, a
feature to suggest, or a change you'd like to make to the code, please open
a new issue or submit a pull request.
Contributors:
Code:
- Cameron Armstrong (sacredbanana/Nightfox)
https://github.com/sacredbanana/
Art:
- Mauricio Sandoval - Icon design
Translations:
- Mauricio Sandoval - Spanish
- Tobias Baeumer - German
Special Thanks:
- Bebbo for creating the Amiga GCC toolchain https://github.com/bebbo
- OpenAI for making this all possible https://openai.com
- EAB and everyone in it for answering my questions https://eab.abime.net/
- J�n Zahuran?��k for all the thorough testing, bundling AmigaGPT into
AmiKit and for all the moral support https://www.amikit.amiga.sk
- CoffinOS for bundling AmigaGPT into CoffinOS https://getcoffin.net
- Amiga Future Magazine for reviewing AmigaGPT and publishing several of
its updates in the News from Aminet section https://www.amigafuture.de/
- WhatIFF? Magaine for reviewing AmigaGPT and interviewing me in issue 14
https://www.whatiff.info
- Dan Wood for reviewing AmigaGPT on his YouTube channel
https://www.youtube.com/watch?v=-OA28r8Up5U
- Proteque-CBN for reviewing AmigaGPT on his YouTube channel
https://www.youtube.com/watch?v=t3q8HQ6wrnw
- AmigaBill for covering AmigaGPT in the Amiga News section on his Twitch
streams and allowing me to join his stream to promote it
https://www.twitch.tv/amigabill
- Les Docs for making a video review and giving a tutorial on how to add
support for the French accent https://www.youtube.com/watch?v=BV5Fq1PresE
- amiga-news.de for covering AmigaGPT news
http://amiga-news.de
Changelog:
------------------
3.0.0 (2026-07-01)
------------------
- Profile settings: New panels to manage chat, image and speech providers
- Chat and Image menus replace the single OpenAI menu; model selection is now
in "Chat Provider Settings" and "Image Provider Settings"
- Fetch latest models available from the server
- System prompt setting is now a chat profile setting rather than a global one
- New Speech Provider Settings window for setting your speech provider and
options, creating speech profiles and testing your current speech settings
- New config options for the Workbench speech systems: rate, pitch, mode
(natural/robotic) and sex
- ARexx SENDMESSAGE, CREATEIMAGE and SPEAKTEXT support PR=PROFILE to choose
profile; new LISTPROFILES command
- ARexx SENDMESSAGE supports SF=SYSTEMFILE to use the contents of a file as the
system prompt
- New powerful Shell Tool option to ask ChatGPT to run any command on
your system in the shell
- The daemon (AmigaGPTD) now maintains conversation history in T: for context
across ARexx calls until reboot
- The installer is now compatible with AmiKit
- Speech now runs in the background so the app can remain responsive while
it's speaking
- Added a button to stop speaking
- ARexx SPEAKTEXT and LISTVOICES support voices for all speech profiles
- New NEWCHAT ARexx command to clear daemon conversation history
- Support for Anthropic Claude for chat
- Support for Anthropic Claude for image generation
- Support for xAI Grok for chat
- Support for xAI Grok for image generation
- Support for xAI Grok for Text-To-Speech
- Added streaming support in chat for Google Gemini
- Support for Google Gemini for image generation
- Automatically instructs the chat server to not use unicode characters the
Amiga cannot display such as emoji since these characters display in
AmigaGPT as question marks
- All config options are now immediately saved to disk
- Added NEWCHAT and PROFILE=PR options for AskGPT
- ARexx say.rexx now supports different profiles and voices
- The returned chat response string in ARexx will no longer contain
unnecessary characters such as escape backslashes and escaped markdown
characters
- Removed the maximum size limit for conversations. Chat max length is now
only determined by available memory and the model's context window
- Greatly improved performance and memory usage when sending chat messages
in the Responses API
- Fix bug where conversation titles were getting displayed in the incorrect
encoding
- Fix bug where the app would get stuck after sending a few messages to Gemini
- Fix bug where say.rexx did not ask you what you want to say
- Fix bug where the first paragraph in every response is always black even if
you have custom colours set
- Fix bug where a missing or incorrect API key could cause the app to enter an
infinite loop of displaying an error and retrying
- Correctly handle reponses that contain JSON in the response header such as
those received from Cloudfare
-------------------
2.17.0 (2025-12-22)
-------------------
- Add support for Elevenlabs TTS
- Cut, Copy, Paste, Select All and Clear commands now work for the text
input box in image generation mode
- Chat system now works for custom LLM servers using the completions endpoint
-------------------
2.16.0 (2025-12-18)
-------------------
- Add GPT Image 1 and GPT Image 1 Mini image models
- Add ability to select image output format (JPG/PNG)
- Deleting images actually now properly deletes them from disk
-------------------
2.15.0 (2025-12-13)
-------------------
- Add support for setting custom provider endpoint and endpoint URL
- Add chat/completions endpoint support for custom providers for compatibility
with servers such as Google Gemini
-------------------
2.14.0 (2025-12-13)
-------------------
- Add GPT 5.2 and GPT 5.2 Pro chat models
-------------------
2.13.2 (2025-12-06)
-------------------
- Chat now scrolls to the bottom instead of the top when you send a chat
message
-------------------
2.13.1 (2025-11-30)
-------------------
- Fix AskGPT crashing when running it without a GUI
- Fix crash in ARexx SENDMESSAGE handler if there's an error response from
the server
-------------------
2.13.0 (2025-11-29)
-------------------
- Add support for custom chat servers that conform to the OpenAI API so you
can chat with alternative LLM servers hosted in your local network or on
the internet
- Add ability for ChatGPT to search the web (enabled by default)
- Add "AskGPT" commandline utility. You can ask AmigaGPT anything from the
shell with the "AskGPT" command
- Add support for the GPT-5.1 chat model
- Fix crash in joke.rexx when asking for another joke
- Fix bug where the new conversation is not the selected conversation in
the conversation list
- Fix bug where starting a new conversation after deleting a conversation
would briefly show the old conversation in the chat window before it gets a response from the server
- Fix freeze when starting a new conversation after the previous conversation
ended in an error
-------------------
2.12.0 (2025-09-22)
-------------------
- Add "verse" OpenAI voice
- Add ability for ARexx command "SPEAKTEXT" to save audio to file of various
audio formats
- The say.rexx script now asks if you want to play the audio or save to a file
with audio format selection
- Improve display of Text To Speech errors
- Most ARexx parameters now use sensible defaults instead of reading from
AmigaGPT config
- Fix bug where conversation titles and ARexx SENDMESSAGE responses were
enclosed in quotes
-------------------
2.11.2 (2025-09-11)
-------------------
- Improve handling of server disconnections
- Fix occasional crash downloading chat responses
- Startup script for the daemon in AmigaOS 4 and MorphOS no longer cause
errors
- The say.rexx script now works with the daemon
-------------------
2.11.1 (2025-09-10)
-------------------
- Can now quit the app and daemon with CTRL-C in the shell
- AmigaGPTD now shows errors in requesters
- AmigaGPTD now launches with sufficient stack size
- Launching ARexx scripts from AmigaGPT now finds RX on earlier versions of
AmigaOS 3
- Say.rexx script now lets you select a voice
-------------------
2.11.0 (2025-09-09)
-------------------
- New daemon app (AmigaGPTD). An app that has no GUI and can run in the
background. If you have this running, there is no longer any need to have
the main AmigaGPT app open if you want to use AmigaGPT's commands in ARexx.
The daemon listens on the ARexx port named "AMIGAGPTD" so make sure your
scripts try to connect to this new port first, and if it fails, attempt
connecting to "AMIGAGPT" which is still the port used by the main app
- Chat message responses from ARexx now put escape sequences in the returned
string instead of using control characters (e.g. \n instead of 0x0A) to make
them easier to work with in ARexx scripts
- Automatically retry sending or receiving data if the request was made before
the handshake had completed
- Greatly improved error reporting for connection errors. It now can show the
user a much more detailed and useful error message if something goes wrong
- Added a 3 minute timeout for waiting for a response from the API so it will
give up after 3 minutes with an error message instead of waiting forever and
making AmigaGPT permanently stuck
- Stop the ARexx scripts that delete a temporary file from announcing the
deletion to the user in the console
- Change the default program for the readme and the guide to both be MultiView
- Installer now installs the updated and new ARexx scripts if an older version
of AmigaGPT is already installed
-------------------
2.10.0 (2025-08-17)
-------------------
- Added menu option to print the conversation with the printer
- Now uses native MUI screen management (set screen in Screen or System menu
in MUI settings)
- Added menu option to clear the MUI settings
- Added more entries to PATH in the ARexx execution environment so ARexx
can run more commands just from their name
- Improve error handling in chat responses
- Restore backwards compatibility with AmigaOS 3.1
- Enter key to submit text now works in MUI 3.9 (not earlier)
- Fix crash opening the "About AmigaGPT" window on MUI 3.9
------------------
2.9.0 (2025-08-09)
------------------
- Added gpt-5 chat models and make this the default chat model
- Fixed bad characters in About AmigaGPT window
- Fix crash switching to fixed width fonts
------------------
2.8.1 (2025-08-04)
------------------
- Fixed some menu items not working
- The window position and size is now remembered between sessions
- Fixed license label in the About AmigaGPT window
------------------
2.8.0 (2025-08-02)
------------------
- Add "Fixed width fonts" item in the "View" menu to toggle fixed width fonts
for the chat input and chat output
- Add "User text alignment" and "Assistant text alignment" items in the "View"
menu to set the alignment of user and assistant messages in the chat
------------------
2.7.0 (2025-07-27)
------------------
- Stop the stack size warning appearing on MorphOS
- Running ARexx scripts now working in MorphOS (requires the 68k version of
rexxsyslib.library installed)
- Fix bug where the conversation window will say "(null)" if the user clicks
in the conversation list below all the entries
- Chat window scrolls to bottom when new text is received
- Can now send the message by pressing return (right shift + return will
create a new line without sending) (MUI 5 only)
- The status bar now has a dark background to make the status more visible
- Fixed bug where your sent message does not get formatted correctly if it
has more than 1 line
- While receiving a message, the chat window only shows the current message.
This dramatically speeds up text rendering. Long conversations no longer
become increasingly sluggish after every new message
------------------
2.6.0 (2025-05-06)
------------------
- Added GPT Image 1 image model
- Added GPT-4.1 mini and GPT-4.1 nano chat models
- Retrieves image data directly from OpenAI instead of downloading from
file server to speed up image retrieval
- Error messages can now be displayed in a requester if MUI is not finished
initialising the app
- Now displays an informative error message if you are attempting to generate
an image when you have insufficient credits in your OpenAI API account
- M68k version compiled with extra optimisations for faster code
- Fixed bug where AmigaGPT would get stuck after sending multiple chat
messages
- Fixed the bsdsocket.library error that appears on MorphOS if the app is
closed before it has finished loading
- Fix crash when toggling speech enabled in menu
- Fix crash when generating an image when OpenAI reports an error
------------------
2.5.0 (2025-04-25)
------------------
- Added an ARexx menu to allow you to run your own scripts. ARexx is not yet
implemented in MorphOS so you can try installing the 68k version of ARexx
and see if that works.
- Added sample scripts
- Status bar background colour adjusted to improve readability
- Latest OpenAI chat models added (gpt-4.1, o3, o4-mini)
- The menu has been rewritten and now gets built dynamically. It will now
automatically create menu options for all the available models. This will
make adding new models much easier and quicker
------------------
2.4.0 (2025-04-05)
------------------
- ARexx port added to allow issuing AmigaGPT commands from another app or from a script
- Fixed bug where the About AmigaGPT window wouldn't display all the text correctly
- Latest OpenAI chat models added
- Latest OpenAI text to speech models added
- Latest OpenAI text to speech voices added
- Can now give instructions to the OpenAI voice to change how it sounds (only works for the gpt-4o-mini-tts model)
- Chat now uses the new Responses endpoint instead of the Chat Completions one to allow for future expansion
------------------
2.3.1 (2025-03-30)
------------------
- Fix broken error requester in MorphOS
- Fix incorrect language name for english-british for OS4
- Running the "Version" command on the AmigaGPT executable will now report the version and build date
- The installer in MorphOS will now properly ask where you want to install the readme
------------------
2.3.0 (2025-03-23)
------------------
- New installer for easy and customisable installation
- Localisation support for 22 languages
- Removed close gadgets from API Key and Chat System requester windows
- Chat system requester text entry field is now multiline and supports clipboard
------------------
2.2.0 (2025-01-04)
------------------
- Support for ApolloOS
- Image preview now works in MorphOS and in MUI 3 (AmigaGPT now has
MCC_Guigfx as a prerequisite)
- Replace the chat output sections from MCC_TextEditor class to
MCC_NFloattext for improved memory usage and compatibility with ApolloOS
- Improved encoding to and from UTF-8 with enhanced character lookalike
remapping (AmigaGPT now has codesets.library as a prerequisite)
- Fix crash when the first data chunk received from OpenAI is smaller than
the HTTP header
- Fix crash when OpenAI sends a blank content string
------------------
2.1.2 (2024-12-31)
------------------
- Fix bug where not all messages were being saved in a new conversation
- Fix menu items not being selectable
------------------
2.1.1 (2024-12-30)
------------------
- Add more clear error message when the user's installed AmiSSL version is too
old
------------------
2.1.0 (2024-12-28)
------------------
- Add proxy server support
------------------
2.0.2 (2024-12-24)
------------------
- Fix memory leaks
- Make more resiliant to connection errors or disconnections
------------------
2.0.1 (2024-12-23)
------------------
- Fix a bug preventing you being able to create a new image after making one
- Fix crash when receiving an error generating an image
------------------
2.0.0 (2024-12-22)
------------------
- Completely replace ReAction with MUI 5 with backwards compatibility with
MUI 3
- Now fully compatible with AmigaOS 3.1
- New native version for MorphOS
- Enhanced "About AmigaGPT" panel
- New "About MUI" panel
- Fix memory leak when receiving messages
- View instant preview of generated images (MUI 5 only and not supported in
MorphOS)
- Updated to the latest OpenAI models
------------------
1.6.3 (2024-09-18)
------------------
- Fixed error receiving chat messages
------------------
1.6.2 (2024-09-01)
------------------
- Increase max OpenAI API key size to 256 characters
------------------
1.6.1 (2024-08-31)
------------------
- Default to speech disabled
- Fix crash when trying to close speech system when it failed to initialise
------------------
1.6.0 (2024-08-10)
------------------
- Add the GPT4o Mini model
- Remove the unused Accent menu item from OS4 version
- Fix bug where some voice menu items in the OS4 version fail to display a
checkmark after you select it
------------------
1.5.1 (2024-06-26)
------------------
- Send button no longer appears after image creation
- Create image button stays disabled after image creation. Select new image to
create a new one
- Updated to latest json-c
------------------
1.5.0 (2024-05-18)
------------------
- Support for OpenAI's 16-bit text to speech voices. Fast internet connection
recommended
- Updated to latest OpenAI chat models (including GPT-4o which will only
work for ChatGPT Plus subscribers for now)
- Removed debug symbols for faster and smaller executable
------------------
1.4.6 (2024-04-13)
------------------
- Send button no longer appears in top left corner after closing a generated
image for Amiga 3.X
------------------
1.4.5 (2024-04-06)
------------------
- Fixes the stack size calculation so it will no longer warn you that the
stack size is too small if it really isn't
------------------
1.4.4 (2024-03-23)
------------------
- Fixes a bug in chat mode in AmigaOS 3.X where the send button stays disabled
after an error occurs
- Change the stack size in the program .info file to the recommended 32768
bytes
- Fix bug where accents were not loaded properly resulting in the speech to
fail
------------------
1.4.3 (2024-02-08)
------------------
- Updated to the latest OpenAI chat models
- Fix crash when config.json doesn't exist
- Fix crash when selecting the root menu items
- Adjust screen colours to enhance visibility
- Improved error handling for connection errors
- Use a stack cookie to set minimum stack size to 32768 bytes (AmigaOS 3.1.4
or higher required)
- Shows a warning if the stack size is smaller than 32768 bytes (AmigaOS 3
only)
- Send button no longer appears in corner of screen on image mode in 3.X after
an error message is dismissed
------------------
1.4.2 (2024-01-31)
------------------
- Replaced clicktabs for AmigaOS 3.X since the version is too old and will
crash. 3.X users can select the mode at startup.
------------------
1.4.1 (2024-01-22)
------------------
- Fixed bug where the prompt textbox remains in a readonly state when
switching from image mode to chat mode
------------------
1.4.0 (2024-01-21)
------------------
- AI image generation! Switch between chat and image generation mode by
clicking the tabs at the top of the screen
- Colour tweaks
- Status bar shows more information and colours
- About screen can now be dismissed with any key press or mouse click
- The OpenAI API key requester now populates with the existing key if it
exists so you can more easily find and fix typos in the key
- Added the ability to set the chat system (new "Chat System" menu item)
- Fix crash when having a conversation consisting of a few messages
------------------
1.3.1 (2023-12-02)
------------------
- Fix pasting into chat input textbox (AmigaOS 3.2 and 4.1 only for now)
- Remove trailing newline from user messages
- Enclose all user lines in asterisks to make bold
- Fix memory leak that was present in downloading responses from OpenAI
------------------
1.3.0 (2023-11-12)
------------------
- Native PPC version for AmigaOS 4.1
- Scrollbar works again
- Improve response handling from OpenAI preventing a crash
- Latest OpenAI models added
------------------
1.2.1 (2023-11-04)
------------------
- Now fully backwards compatible with Cloanto's AmigaOS 3.X (included
in Amiga Forever)
------------------
1.2.0 (2023-10-04)
------------------
- Now fully backwards compatible with AmigaOS 3.9
------------------
1.1.0 (2023-07-22)
------------------
- Now supports characters from Western languages other than English
- Blank responses from OpenAI no longer cause a crash
------------------
1.0.1 (2023-06-24)
------------------
- After setting UI font, the font is now applied fully without needing
to restart the app
----------------
1.0 (2023-06-17)
----------------
- Initial release
License:
--------
AmigaGPT
--------
MIT License
Copyright (c) 2023-2025 Cameron Armstrong
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
--------------------
THIRD PARTY LICENSES
--------------------
------
json-c
------
Copyright (c) 2009-2012 Eric Haszlakiewicz
Permission is hereby granted, free of charge, to any person obtaining a
copy of this software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation
the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included
in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
----------------------------------------------------------------
Copyright (c) 2004, 2005 Metaparadigm Pte Ltd
Permission is hereby granted, free of charge, to any person obtaining a
copy of this software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation
the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included
in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
-----
Flite
-----
Flite is free software.
We have kept the core code to BSD-like copyright, thus the system is
free to use in commercial products, with commercial extensions. GPL
code is only included as part of the build process and does not
taint any of the run-time code.
Note the PalmOS code uses BSD code that requires acknowledgements in
any distribution documentation.
As a collection it is distributed under the following license. Note
a few files in this distribution have a different but equally free
non-conflicting licence, see below.
Language Technologies Institute
Carnegie Mellon University
Copyright (c) 1999-2008
All Rights Reserved.
Permission is hereby granted, free of charge, to use and distribute
this software and its documentation without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of this work, and to
permit persons to whom this work is furnished to do so, subject to
the following conditions:
1. The code must retain the above copyright notice, this list of
conditions and the following disclaimer.
2. Any modifications must be clearly marked as such.
3. Original authors' names are not deleted.
4. The authors' names are not used to endorse or promote products
derived from this software without specific prior written
permission.
CARNEGIE MELLON UNIVERSITY AND THE CONTRIBUTORS TO THIS WORK
DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING
ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT
SHALL CARNEGIE MELLON UNIVERSITY NOR THE CONTRIBUTORS BE LIABLE
FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
THIS SOFTWARE.
All files within this distribution have the above license except
the following
src/cg/cst_mlpg.h
src/cg/cst_mlpg.c
src/cg/cst_mlsa.h
src/cg/cst_mlsa.c
src/cg/cst_vc.h
src/cg/cst_vc.c
/*********************************************************************/
/* */
/* Nagoya Institute of Technology, Aichi, Japan, */
/* Nara Institute of Science and Technology, Nara, Japan */
/* and */
/* Carnegie Mellon University, Pittsburgh, PA */
/* Copyright (c) 2003-2004 */
/* All Rights Reserved. */
/* */
/* Permission is hereby granted, free of charge, to use and */
/* distribute this software and its documentation without */
/* restriction, including without limitation the rights to use, */
/* copy, modify, merge, publish, distribute, sublicense, and/or */
/* sell copies of this work, and to permit persons to whom this */
/* work is furnished to do so, subject to the following conditions: */
/* */
/* 1. The code must retain the above copyright notice, this list */
/* of conditions and the following disclaimer. */
/* 2. Any modifications must be clearly marked as such. */
/* 3. Original authors' names are not deleted. */
/* */
/* NAGOYA INSTITUTE OF TECHNOLOGY, NARA INSTITUTE OF SCIENCE AND */
/* TECHNOLOGY, CARNEGIE MELLON UNIVERSITY, AND THE CONTRIBUTORS TO */
/* THIS WORK DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, */
/* INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, */
/* IN NO EVENT SHALL NAGOYA INSTITUTE OF TECHNOLOGY, NARA */
/* INSTITUTE OF SCIENCE AND TECHNOLOGY, CARNEGIE MELLON UNIVERSITY, */
/* NOR THE CONTRIBUTORS BE LIABLE FOR ANY SPECIAL, INDIRECT OR */
/* CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM */
/* LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, */
/* NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN */
/* CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */
/* */
/*********************************************************************/
These functions are derived from the versions in festvox/src/vc/ as
modified by Tomoki Toda which in turn contain code derived from
NITECH's HTS system. Their copyright has the same freedoms as
as Flite's but under NAIST, NITECH and/or CMU.
src/audio/au_wince.c
src/utils/cst_file_stdio.c
src/utils/cst_mmap_posix.c
src/utils/cst_mmap_win32.c
src/utils/cst_mmap_none.c
src/utils/cst_file_wince.c
sapi/
are copyright Cepstral, LLC rather than CMU but fall under the same
free licence as the above, except for the owner.
doc/alice
Is the first two chapters of Alice in Wonderland as distributed by the
Gutenburg project and is now in the public domain
src/regex/regexp.c
src/regex/regsub.c
* Copyright (c) 1986 by University of Toronto.
* Written by Henry Spencer. Not derived from licensed software.
*
* Permission is granted to anyone to use this software for any
* purpose on any computer system, and to redistribute it freely,
* subject to the following restrictions:
*
* 1. The author is not responsible for the consequences of use of
* this software, no matter how awful, even if they arise
* from defects in it.
*
* 2. The origin of this software must not be misrepresented, either
* by explicit claim or by omission.
*
* 3. Altered versions must be plainly marked as such, and must not
* be misrepresented as being the original software.
src/speech/rateconv.c
* Copyright (c) 1992, 1995 by Markus Mummert
*
* Redistribution and use of this software, modifcation and inclusion
* into other forms of software are permitted provided that the following
* conditions are met:
*
* 1. Redistributions of this software must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. If this software is redistributed in a modified condition
* it must reveal clearly that it has been modified.
lang/usenglish/us_durz_cart.c
lang/usenglish/us_durz_cart.h
lang/usenglish/us_int_accent_cart.c
lang/usenglish/us_int_accent_cart.h
lang/usenglish/us_int_tone_cart.c
lang/usenglish/us_int_tone_cart.h
lang/usenglish/us_phoneset.c
lang/usenglish/us_f0lr.c
These are directly (or indirectly) compiled/derived from files that are
part of the Festival Speech Synthesis System (1.4.1). Hence they have
a joint copyright CMU/Edinburgh but with the same free licence
configure
# Copyright (C) 1992, 93, 94, 95, 96 Free Software Foundation, Inc.
#
# This configure script is free software; the Free Software Foundation
# gives unlimited permission to copy, distribute and modify it.
configure.sub
config.guess
missing
install-sh
mkinstalldirs
Copyright FSF, and under the GPL, these files are only used for
convenient configuration and are not part of the generated binary,
and therefore do not impose any GPL restrctions on the rest of the
system. But as they are standard methods for configuration they
are included.
palm/include/peal.h
palm/include/pealstub.h
palm/arm_flite/pealstub.c
palm/m68k_flite/peal.c
/**********
* Copyright (c) 2004 Greg Parker. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY GREG PARKER ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
**********/
The whole Sun Microsystems math library was ported to Palm
for this but its not actually needed, though the include file is
to allow bits of the system to compile that currently linked in
to the system, so we just have the include file here
palm/include/fdlibm.h
* ====================================================
* Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
*
* Developed at SunSoft, a Sun Microsystems, Inc. business.
* Permission to use, copy, modify, and distribute this
* software is freely granted, provided that this notice
* is preserved.
* ====================================================
------------
Flite Device
------------
/* Copyright (c) 2013 Fredrik Wikstrom. All rights reserved.
**
** Redistribution and use in source and binary forms, with or without
** modification, are permitted provided that the following conditions
** are met:
**
** 1. Redistributions of source code must retain the above copyright
** notice, this list of conditions and the following disclaimer.
**
** 2. Redistributions in binary form must reproduce the above copyright
** notice, this list of conditions and the following disclaimer in the
** documentation and/or other materials provided with the distribution.
**
** THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS `AS IS'
** AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
** IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
** ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
** LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
** CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
** SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
** INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
** CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
** ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
** POSSIBILITY OF SUCH DAMAGE.
*/
|
aminet net>