AmigaGPT 3.0.0 released with Claude, Gemini and Grok support for AmigaOS

AmigaGPT 3.0.0 is not a cosmetic point release. It is a structural update that turns the application from a primarily OpenAI-focused Amiga chatbot into a broader AI front end for several major model providers, while also tightening automation, speech handling, configuration and day-to-day responsiveness. For AmigaOS 3.1, AmigaOS 4.1 and MorphOS users, the headline change is clear: AmigaGPT now reaches beyond ChatGPT and adds practical support for Claude, Gemini and Grok across key parts of the program. The result is a more flexible AI client for Amiga systems, with provider selection moved into a cleaner profile-based model and several important changes designed for users who rely on ARexx, shell workflows, speech output or long conversations. Version 3.0.0 does not try to reinvent the interface for its own sake. Instead, it reorganises the application around the way AI services are now actually used: different models for different tasks, different providers for different capabilities, and more control over how each profile behaves.

A wider choice of AI providers

The most important change in AmigaGPT is the expansion of supported AI services. The application now adds support for Anthropic Claude, xAI Grok and Google Gemini in addition to its existing OpenAI functionality. This matters because model choice is no longer a minor preference. Some users want fast chat responses, others need stronger reasoning, others want image generation, and others may already have API access with a particular provider.

Claude support arrives for chat and image generation, while Grok support extends to chat, image generation and text-to-speech. Gemini support is also expanded, including image generation and streaming chat support. That last point is especially relevant on older or constrained systems, because streaming output makes interaction feel more immediate instead of forcing the user to wait for a full reply before anything appears.

This broader provider support gives AmigaGPT a more modern footing without turning the program into a bloated launcher. It remains an Amiga-native client, but it now acknowledges the current AI landscape, where users often move between models depending on cost, speed, quality and availability.

Profiles replace the old single-provider mindset

Version 3.0.0 also changes how settings are organised. New profile settings panels now manage chat, image and speech providers, replacing the previous single OpenAI menu approach with separate Chat and Image menus. Model selection has moved into dedicated Chat Provider Settings and Image Provider Settings windows, making the configuration more logical and easier to scale as more services are added.

The system prompt is now handled as a chat profile setting rather than a global option. This is a sensible change because system prompts are often task-specific. A user might want one profile for general chat, another for coding help, another for concise command-line use and another for a highly controlled automation workflow. Keeping the system prompt inside the profile prevents one use case from accidentally bleeding into another.

AmigaGPT can also fetch the latest models available from the server, reducing the need for users to manually track model names or update configuration details whenever a provider changes its catalogue. In a software category where available models shift frequently, this is a practical quality-of-life improvement.

Speech handling becomes more capable

Speech support receives a substantial refresh in AmigaGPT. A new Speech Provider Settings window allows users to choose a speech provider, adjust options, create speech profiles and test the current speech setup. For Workbench speech systems, new configuration options include rate, pitch, mode and sex, giving users more control over output style.

The application also becomes more responsive while speaking. Speech now runs in the background, so the main program no longer has to feel blocked while audio output is in progress. A new stop speaking button adds another basic but important usability improvement, particularly when long responses are being read aloud or the user simply wants to interrupt output.

ARexx and shell workflows get serious attention

AmigaGPT has already been notable for its ARexx integration, and version 3.0.0 pushes that further.  The companion daemon, AmigaGPTD, now maintains conversation history in T: for context across ARexx calls until reboot. This improves scripted conversations because separate calls no longer have to behave like isolated one-shot requests. A new new chat ARexx command clears daemon conversation history when a fresh context is needed.

There is also a powerful new Shell Tool option that can ask ChatGPT to run commands on the system shell. This is one of the most consequential additions in the update and also one that users should approach with care. Used sensibly, it could speed up command-line tasks and administrative routines. Used casually, it has the same risk profile as any tool that allows AI-generated commands to affect a live system.

Longer chats and better performance

AmigaGPT removes the previous maximum size limit for conversations. Chat length is now determined by available memory and the model’s context window rather than an artificial ceiling inside the application. For users working through longer technical explanations, scripts, documents or multi-step troubleshooting, this is a meaningful change.

Performance and memory usage have also been improved when sending chat messages through the Responses API. This is not the sort of feature that dominates a changelog headline, but it may be one of the changes users notice most in regular use. AI clients can become sluggish quickly when conversation history grows, and efficiency matters even more on Amiga-class hardware.

Configuration handling has also been improved, with all config options now saved to disk immediately. That reduces the chance of losing changes after a crash, reset or unexpected exit, and it makes the overall settings experience feel more dependable.

Cleaner output and important bug fixes

Several fixes in 3.0.0 target the kind of rough edges that affect real daily use. The update fixes an issue where conversation titles were displayed with incorrect encoding, and it addresses a Gemini-related problem where the application could become stuck after a few messages. It also fixes a problem where the first paragraph in every response appeared black even when custom colours were set.

Error handling has been improved as well. A missing or incorrect API key should no longer trigger an infinite loop of repeated error messages and retries. Responses containing JSON in the response header, including those received through Cloudflare, are now handled correctly. ARexx output is also cleaner, with returned chat response strings no longer containing unnecessary escape backslashes and escaped Markdown characters.

The installer is now compatible with AmiKit, widening the number of prepared environments where installation should be smoother. AskGPT also gains NEWCHAT and PROFILE options, while the say.rexx script now supports different profiles and voices.

A focused update with real workflow impact

AmigaGPT 3.0.0 is a significant release because it updates the application around the current reality of AI use. The move to multiple providers, profile-based configuration, stronger speech controls, expanded ARexx commands and improved daemon context makes the program more flexible without losing its Amiga identity.

For users who only want a simple AI chat window, the biggest benefit is choice: ChatGPT, Claude, Gemini and Grok can now sit within the same broader workflow. For power users, the more important story is control. Profiles, scriptable provider selection, file-based system prompts, persistent daemon context and improved shell integration make AmigaGPT a more capable tool for automation as well as conversation.

This is the kind of update that changes how a program can be used, not just how it looks. AmigaGPT turns multi-provider AI, speech output and scripted interaction into more coherent parts of one application, making it one of the more substantial updates yet for AI access on AmigaOS and MorphOS.