v2.10

This update brings improvements to the Translate command.

We now support 59 languages, and they are all the languages currently supported by ChatGPT.

Additionally, the command now defaults to using the previously selected language, so you don't have to choose your preferred language every time!

Full Changelog: v2.9...v2.10

v2.9

Added the DuckDuckGo provider. It supports the following models:

• gpt-4o-mini
• claude-3-haiku
• meta-llama-3-70b
• mixtral-8x7b

two of which are new.

DuckDuckGo also has a very strong privacy guarantee.

What's Changed

• Add DuckDuckGo provider by @XInTheDark in #88

Full Changelog: v2.8...v2.9

v2.8

In this update:

• Added the SambaNova provider. It offers extremely fast Llama-3.1 405B, 70B and 8B models (up to 1000 tokens/sec during testing).
• Enhancements for Local API configuration.

What's Changed

• Add SambaNova provider by @XInTheDark in #86

Full Changelog: v2.7.1...v2.8

v2.7.1

Added more Local API config settings. You can now change the Ollama API URL, as well as the context size during inference.

What's Changed

• Add more Ollama config settings by @XInTheDark in #85

Full Changelog: v2.7...v2.7.1

v2.7

In this update:

• A new model llava-1.5-7b was added. It is capable of understanding images, similar to Gemini. Simply upload files and chat with it!
• Compressed asset files (extension and command icons). There has been no decrease in quality, and the total file size has decreased from 2.6MB to less than 800KB!
• Cosmetic changes: updated the README with latest info, and added screenshots.

What's Changed

• Compress asset files by @XInTheDark in #80
• Add llava-1.5-7b model (hosted by DeepInfra) by @XInTheDark in #81
• Add screenshots to README by @XInTheDark in #82

Full Changelog: v2.6...v2.7

v2.6

In this update:

• Added the Llama 3.1 405B model to Replicate provider.
• Added the Meta AI provider, powered by the latest Llama 3.1 models, and with internet access.
• Added the Edit Chat Settings action in AI Chat, allowing you to change settings even after chat creation.
• (Minor cosmetic change) Changed Toast messages to have a consistent capitalization style.
• Code cleanups.

What's Changed

• Add Llama 3.1 405B (hosted by Replicate) by @XInTheDark in #76
• Add Meta AI provider by @XInTheDark in #77
• Add "Edit Chat Settings" action in AI Chat, change Toast messages by @XInTheDark in #79

Full Changelog: v2.5...v2.6

v2.5

We now support local inference with Ollama! Ollama lets you run various open-source LLMs locally, in a Local API. To use local inference, you need to have Ollama installed.

If you're interested, learn how to get started here! https://github.com/XInTheDark/raycast-g4f/wiki/Help-page:-Configure-Local-APIs#getting-started-ollama

In this update:

• Added the "Ollama Local API" provider.
• The "Configure GPT4Free Local API" command has been renamed to "Configure Local APIs", and it now manages the settings for both G4F and Ollama APIs.

What's Changed

• Fix Blackbox formatting by @XInTheDark in #72
• Add Ollama Local API support by @XInTheDark in #74

Full Changelog: v2.4...v2.5

v2.4

Llama 3.1 is out! The 405B, 70B and 8B models have all been added to this extension.
Also deprecated the non-functional Ecosia provider.

What's Changed

• Remove Ecosia provider by @XInTheDark in #69
• Add Llama 3.1 models (hosted by DeepInfra) by @XInTheDark in #71

Full Changelog: v2.3...v2.4

v2.3

This update brings a new feature: AI Presets!

AI Presets allow you to create your own chat templates. You can create new presets and specify the model, system instructions and creativity settings. Your presets can then be used in AI Chat, so you don't have to enter the same settings for every new chat.

Read more about how to use them here!

What's Changed

• Add AI Presets, "Manage AI Presets" command by @XInTheDark in #68

Full Changelog: v2.2.1...v2.3

v2.2.1

This is a bugfix update. The following issue was resolved:

• Recent updates to the DeepInfra provider caused the API results to become incompatible. The extension has been patched accordingly.

Full Changelog: v2.2...v2.2.1